KR100843820B1 - Information signal processing device, information signal progessing method, image signal processing device and image display device using it, coefficient type data creating device used therein and creating method, coefficient data creating device and creating method, and information providing medium - Google Patents

Abstract

The present invention relates to, for example, an information signal processing apparatus suitable for application when converting an SD signal into an HD signal. Using the pixel data of the tap corresponding to the target position of the HD signal, selectively extracted from the SD signal, a class CL to which the pixel data of the target position belongs. The coefficient generation circuit 136 generates coefficient data Wi of each class based on coefficient seed data of each class and a plurality of image quality adjustment parameters h and v values by user operation. The arithmetic circuit 127 is provided with the data xi of the tap corresponding to the target position of the HD signal, selectively extracted from the SD signal by the tap selection circuit 121, and the coefficient data corresponding to the class CL read out from the memory 134. By using the estimation equation using Wi, pixel data of the target position of the HD signal is obtained. Image quality can be adjusted freely on a plurality of axes.

SD signal, HD signal, coefficient seed data generation, image quality adjustment parameters

Description

An information signal processing apparatus, an information signal processing method, an image signal processing apparatus, and an image display apparatus using the same, a coefficient seed data generating apparatus and generating method used therein, a coefficient data generating apparatus and a generating method, and an information providing medium {INFORMATION SIGNAL PROCESSING DEVICE, INFORMATION SIGNAL PROGESSING METHOD, IMAGE SIGNAL PROCESSING DEVICE AND IMAGE DISPLAY DEVICE USING IT, COEFFICIENT TYPE DATA CREATING DEVICE USED THEREIN AND CREATING METHOD, COEFFICIENT DATA CREATING DEVICE AND CREATING METHOD, AND INFORMATION PROVIDING MEDIUM}

The present invention relates to an information signal processing apparatus and an information signal processing method which perform, for example, function processing such as resolution enhancement, noise suppression, decoding, signal format conversion, and the like.

Specifically, in response to a parameter value for selecting one function among a plurality of functions, a plurality of function processing is performed by a single device by generating the information data constituting the second information signal from the information data constituting the first information signal. The present invention relates to an information signal processing apparatus and the like which can realize the above.

The present invention is, for example, an information signal processing apparatus, an information signal processing method, an image signal processing apparatus, and an image display apparatus using the same, which are suitable for application when converting an NTSC type video signal into a high-vision video signal. It relates to a coefficient seed data generating device and generating method, a coefficient data generating device and generating method and an information providing medium. Specifically, when converting the first information signal to the second information signal, by generating the second information signal corresponding to a plurality of types of parameter values, the quality of the output obtained by the second information signal, for example, the image quality An information signal processing apparatus or the like for allowing adjustment to be freely performed on a plurality of axes.

In recent years, the development of TV receivers capable of obtaining higher resolution images has been demanded due to the increase in audio visual orientation, and so-called high vision has been developed in response to this demand. The number of scanning lines in Hi-vision is 1125, which is more than twice that of 525 scanning lines in NTSC. In addition, the aspect ratio of the high-vision is 9:16 compared to the aspect ratio of the NTSC system is 3: 4. Therefore, in Hi-vision, a realistic image can be displayed at a higher resolution than the NTSC system.

Although high-vision has such excellent characteristics, high-vision image display cannot be performed even if the NTSC system video signal is supplied as it is. This is because the NTSC system and the high vision have different specifications as described above.

Therefore, in order to display an image suitable for an NTSC video signal in a high-vision system, the present applicant first proposed a conversion device for converting an NTSC-based video signal to a high-vision video signal (Japanese Patent Application No. 6-). 205934). The converter extracts pixel data of a block (area) corresponding to the main position of a high-vision video signal from an NTSC video signal, and based on the level distribution pattern of the pixel data of this block, The class is determined, and pixel data of the target position is generated corresponding to the class.

In addition, the present applicant has the same configuration as the above-described converter, and has a converter (see Japanese Patent Application Laid-Open No. 2000-138950) and a device for decoding MPEG (Moving Picture Experts Group) image signals (converted from a composite signal to a component signal). Japanese Patent Application No. 2000-135356).

In the above-described conversion apparatus for converting an NTSC video signal into a high-vision video signal, the resolution of the image by the high-vision video signal is fixed, and conventionally, such as adjustment of contrast, sharpness, etc. I could not achieve the desired resolution.

In addition, each of the above-described conventional apparatuses is not efficient by performing a single function process.

An object of the present invention is to provide an information signal processing apparatus or the like capable of realizing processing of a plurality of functions by a single apparatus. In addition, an object of the present invention is to provide an information signal processing apparatus or the like that allows, for example, adjustment of image quality of an image to be freely performed on a plurality of axes such as horizontal resolution and vertical resolution, noise rejection, and horizontal / vertical resolution. It is done.

An information signal processing apparatus according to the present invention performs a processing of one function determined from a plurality of functions on an input first information signal to generate a second information signal, and outputs the second information signal. An apparatus, comprising: parameter input means for inputting a parameter value for determining one function among a plurality of functions, and a second information signal from information data constituting the first information signal corresponding to the parameter value input to the parameter input means; Information data generating means for generating the constituent information data.

In addition, the information signal processing method according to the present invention performs the processing of one function determined from a plurality of functions with respect to an input first information signal to generate a second information signal, and outputs the second information signal. A signal processing method comprising: a first step of inputting a parameter value for determining one function among a plurality of functions, and second information from information data constituting the first information signal corresponding to the parameter value input in this first step; And a second step of generating information data constituting the signal.

In the present invention, a first information signal is input, and processing of one function selected from a plurality of functions is performed on the first information signal to generate a second information signal, and the second information signal is output. In this case, a parameter value for determining one of the plurality of functions is input. For example, when the information signal is an image signal, the plurality of functions are resolution enhancement, noise suppression, decoding, signal format conversion, and the like.

Corresponding to the parameter value input in this way, the information data constituting the second information signal is generated from the information data constituting the first information signal. In this way, the function is switched according to the input parameter value. That is, the processing of a plurality of functions can be realized by a single device.

An information signal processing apparatus according to the present invention is an information signal processing apparatus for converting a first information signal composed of a plurality of information data into a second information signal composed of a plurality of information data, the second information being based on the first information signal. First data selecting means for selecting a plurality of first information data located around the target position in the signal and the plurality of first information data selected by the first data selecting means, the information data of the target position By means of class detecting means for detecting the class to which it belongs, parameter adjusting means for adjusting a plurality of types of parameter values for determining the quality of the output obtained by the second information signal, and class and parameter adjusting means detected by the class detecting means. Information data for generating information data of the target position corresponding to the plurality of types of parameter values It comprises generating means.

For example, the information data generating means is configured to store coefficient seed data which is coefficient data in a generation formula including a plurality of types of parameters for generating coefficient data used in the estimation formula, which are obtained in advance for each class detected by the class detecting means. The class and parameter adjustments generated by the generation formula using the first memory means, the coefficient seed data stored in the first memory means and the plurality of types of parameter values adjusted by the parameter adjusting means, and detected by the class detecting means. Coefficient data generating means for generating coefficient data of the estimation equation corresponding to a plurality of types of parameter values adjusted by the means, and a plurality of positions located around the target position in the second information signal based on the first information signal. Second data selecting means for selecting second information data and counting data generating means; Using a plurality of the second information data is selected and the coefficient data generated by the second data selection means, based on it said estimation includes a calculation means for obtaining the output information data of the target position.

Further, for example, the information data generating means has a memory for storing coefficient data of an estimated expression generated in advance for each combination of a class detected by the class detecting means and a plurality of types of parameter values adjusted by the parameter adjusting means, and class detection Coefficient data generating means for generating coefficient data of the estimation equation corresponding to the class value detected by the means and the parameter adjusting means detected by the means, and the position of interest in the second information signal based on the first information signal; The estimation using the second data selecting means for selecting a plurality of second information data located in the periphery, and the coefficient data generated by the coefficient data generating means and the plurality of second information data selected by the second data selecting means. And calculating means for calculating and obtaining the information data of the target position based on the expression.

Further, the information signal processing method according to the present invention is an information signal processing method for converting a first information signal composed of a plurality of information data into a second information signal composed of a plurality of information data, and from the first information signal to the second information. A first step of selecting a plurality of first information data located in the periphery of the target position in the signal, and based on the plurality of first information data selected in the first stage, the class to which the information data of the target position belongs. A second step of detecting, a third step of adjusting a plurality of types of parameter values for determining the quality of output obtained by the second information signal, and a plurality of types of classes adjusted in the second step and a class detected in the second step And a fourth step of generating information data of the location of interest, in response to a parameter value.

The information providing medium according to the present invention also provides a computer program for executing each step of the above-described information signal processing method.

The image signal processing apparatus according to the present invention is an image signal processing apparatus for converting a first image signal composed of a plurality of pixel data into a second image signal composed of a plurality of pixel data. A data selection means for selecting a plurality of pixel data positioned around the target position in the two image signals and a class to which the pixel data of the target position belongs based on the plurality of pixel data selected by the data selection means. A plurality of parameter values adjusted by class detecting means, parameter adjusting means for adjusting plural kinds of parameter values for determining the quality of output obtained by the second image signal, and class and parameter adjusting means detected by the class detecting means. Corresponding to the pixel data generation number for generating the pixel data of the target position Includes only.

In addition, the image display device according to the present invention includes an image signal input means for inputting a first image signal composed of a plurality of pixel data, and a first image signal inputted from the image signal input means composed of a plurality of pixel data. Image signal processing means for converting and outputting two image signals, image display means for displaying an image by a second image signal output from the image signal processing means on an image display element, and an image displayed on the image display element. And parameter adjusting means for adjusting a plurality of types of parameter values for determining image quality. The image signal processing means includes data selection means for selecting a plurality of pixel data located around the target position in the second image signal based on the first image signal, and a plurality of pixel data selected by the data selection means. Based on the class detecting means for detecting the class to which the pixel data of the target position belongs, and the plurality of types of parameter values adjusted by the class and the parameter adjusting means detected by the class detecting means, the pixel of the target position Pixel data generating means for generating data.

In the present invention, a plurality of first information data located around the target position in the second information signal is selected based on the first information signal, and the information of the target position is based on the plurality of first information data. The class to which the data belongs is detected. For example, the level distribution patterns of the plurality of first information data are detected, and the class to which the information data of the target position belongs is detected based on this level distribution pattern. Here, the information signal is an image signal or an audio signal, for example.

With the parameter adjusting means, a plurality of types of parameter values for determining the quality of the output obtained by the second information signal are adjusted. For example, when the information signal is an image signal, the parameter value is adjusted and the image quality of the image by the second information signal (image signal) is determined. In addition, when the information signal is a voice signal, the parameter value is adjusted to determine the sound quality of the voice by the second information signal (voice signal). For example, the parameter adjusting means is configured to have display means for displaying adjustment positions of plural kinds of parameters and user operation means for adjusting plural kinds of parameter values with reference to the display of the display means. Thereby, a user can operate a user operation means, for example a pointing device, and can easily adjust a plurality of types of parameter values to a desired position.

Then, in response to the detected class and the adjusted plural kinds of parameter values, information data of the target position is generated. For example, coefficient seed data, which is coefficient data in a generation formula for generating coefficient data used in a preformed equation obtained for each class, is stored in the memory means, and detected using this coefficient seed data and adjusted parameter values. The coefficient data of the estimation equation corresponding to the class and the adjusted plural kinds of parameter values are generated, and a plurality of second information data located around the target position in the second information signal is selected based on the first information signal. Then, using this coefficient data and the plurality of second information data, the information data of the target position is generated based on the estimation formula.

Further, for example, coefficient data of an estimated equation generated in advance for each combination of class and plural kinds of parameter values is stored in a memory, and from this memory, an estimated equation corresponding to the detected class and adjusted plural kinds of parameter values In addition to reading the coefficient data of, a plurality of second information data located around the target position in the second information signal is selected based on the first information signal, and the coefficient data and the plurality of second information data are used. Thus, the information data of the target position is generated based on the estimation formula.

In this way, coefficient data of the estimation equation corresponding to the adjusted plural kinds of parameter values are obtained, and this coefficient data is used, and information data of the position of interest in the second information signal is generated based on the estimation equation. Therefore, the quality of the output obtained by the second information signal, for example, the image quality, can be freely adjusted in a plurality of axes such as horizontal resolution and vertical resolution, noise reduction and horizontal / vertical resolution.

Furthermore, the coefficient seed data generating device according to the present invention is for generating coefficient data used in an estimation equation used when converting a first information signal composed of a plurality of information data into a second information signal composed of a plurality of information data. An apparatus for generating coefficient seed data, which is coefficient data in a generation expression, comprising: signal processing means for processing a teacher signal corresponding to a second information signal to obtain an input signal corresponding to the first information signal; Parameter adjustment means for adjusting a plurality of types of parameter values for determining the quality of an output obtained by the input signal in correspondence with the plurality of types of parameters included in the; and around the point of interest in the teacher signal based on the input signal. First data selecting means for selecting a plurality of first information data to be located and selected by the first data selecting means Class detecting means for detecting a class to which the information data of the target position belongs based on the plurality of first information data, and a plurality of second information data positioned around the target position in the teacher signal based on the input signal. Coefficient class for each class, using second data selecting means for selecting a, a class detected by the class detecting means, a plurality of second information data selected by the second data selecting means, and information data of the point of interest in the teacher signal. Regular equation generating means for generating a normal equation for obtaining data, and coefficient seed data calculating means for solving the normal equation and obtaining coefficient seed data for each class.                 

Further, the coefficient seed data generating method according to the present invention is for generating coefficient data used in an estimation equation used when converting a first information signal composed of a plurality of information data into a second information signal composed of a plurality of information data. A method of generating coefficient seed data, which is coefficient data in a generation equation, comprising: a first step of processing a teacher signal corresponding to a second information signal to obtain an input signal corresponding to the first information signal; A second step of adjusting a plurality of types of parameter values for determining the quality of the output obtained by the input signal in correspondence with the types of parameters; and based on the input signal, a plurality of positions located around the target position in the teacher signal. A position of interest based on a third step of selecting first information data and a plurality of first information data selected in this third step; A fourth step of detecting a class to which the information data belongs, a fifth step of selecting a plurality of second information data located around the target position in the teacher signal based on the input signal, and the detected in the fourth step A sixth step of generating a regular equation for obtaining coefficient seed data from the class, the plurality of second information data selected in the fifth step, and the information data of the target position in the teacher signal; And a seventh step of obtaining coefficient seed data.

The information providing medium according to the present invention also provides a computer program for executing each step of the above-described method for generating coefficient seed data.

In the present invention, the teacher signal corresponding to the second information signal is processed to obtain an input signal corresponding to the first information signal. In this case, the quality of the output obtained by the input signal is determined by the adjusted plural kinds of parameter values. For example, when the information signal is an image signal, plural kinds of parameter values are adjusted, and the image quality of the image by the input signal is determined. In addition, when the information signal is an audio signal, a plurality of types of parameter values are adjusted to determine the sound quality of the audio by the input signal.

Based on this input signal, a plurality of first information data located around the target position in the teacher signal is selected, and based on the plurality of first information data, a class to which the information data of the target position belongs is detected. do. Further, based on this input signal, a plurality of second information data located around the target position in the teacher signal is selected.

Then, a plurality of types of parameter values are adjusted in a plurality of steps, and counted by class, from the class to which the information data of the target position in the teacher signal belongs, the selected plurality of second information data and the information data of the target position in the teacher signal A normal equation for obtaining data is generated, and by solving this equation, coefficient seed data for each class is obtained.

Here, the coefficient seed data is coefficient data in a generation equation including the above-described plurality of parameters for generating coefficient data used in the estimation equation used when converting from the first information signal to the second information signal. By using this coefficient seed data, it is possible to obtain coefficient data corresponding to a plurality of types of parameter values arbitrarily adjusted by the generation formula. Accordingly, when converting from the first information signal to the second information signal using the estimation equation, the quality of the output obtained by the second information signal can be freely adjusted on the plurality of axes by adjusting a plurality of types of parameter values. have.                 

Furthermore, the coefficient seed data generating device according to the present invention is for generating coefficient data used in an estimation equation used when converting a first information signal composed of a plurality of information data into a second information signal composed of a plurality of information data. An apparatus for generating coefficient seed data, which is coefficient data in a generation expression, comprising: signal processing means for processing a teacher signal corresponding to a second information signal to obtain an input signal corresponding to the first information signal, and a plurality of included in the generation expression; A parameter adjusting means for adjusting a plurality of types of parameter values for determining the quality of output obtained by the input signal in correspondence with the types of parameters, and a plurality of positions located around the point of interest in the teacher signal based on the input signal. First data selecting means for selecting first information data, and a plurality of selected by the first data selecting means Class detecting means for detecting a class to which the information data of the target position belongs based on the first information data, and based on an input signal, selecting a plurality of second information data located around the target position in the teacher signal. By using the second data selecting means, the class detected by the class detecting means, the plurality of second information data selected by the second data selecting means, and the information data of the target position in the teacher signal, the class and the plurality of types of parameter values First normal equation generating means for generating a first normal equation for obtaining coefficient data of the estimated equation for each combination, coefficient data calculating means for solving the first normal equation and obtaining coefficient data of the estimated equation for each combination; Obtain coefficient seed data for each class using coefficient data for each combination obtained by coefficient data calculating means. And a second comprising a second normal equation production means and the second coefficient seed data calculation means for releasing the normal equation, for obtaining the coefficient seed data for each class, generating a normal equation.

A coefficient seed data generation method according to the present invention is a generation equation for generating coefficient data used in an estimation equation used when converting a first information signal composed of a plurality of information data into a second information signal composed of a plurality of information data. A method of generating coefficient seed data, the coefficient data of which comprises: a first step of processing a teacher signal corresponding to a second information signal to obtain an input signal corresponding to the first information signal; A second step of adjusting a plurality of types of parameter values for determining the quality of output obtained by the input signal in correspondence with the parameters; and a plurality of firsts positioned around the target position in the teacher signal based on the input signal. Based on the third step of selecting the information data and the plurality of first information data selected in the third step, the position of interest is determined. A fourth step of detecting a class to which the data belongs, a fifth step of selecting a plurality of second information data located around the target position in the teacher signal based on the input signal, and the class detected in the fourth step Generating a first normal equation for obtaining coefficient data of the estimated equation for each combination of class and plural kinds of parameter values from the plurality of second information data selected in the fifth step and the information data of the target position in the teacher signal. Using the sixth step, solving the first normal equation, obtaining the coefficient data of the estimation formula for each of the combinations, and using the coefficient data for each combination obtained in this seventh step, the coefficient seed data is obtained for each class. And an eighth step of generating a second normal equation for obtaining, and a ninth step of solving the second normal equation and obtaining coefficient seed data for each class.                 

The information providing medium according to the present invention also provides a computer program for executing each step of the above-described method for generating coefficient seed data.

In the present invention, the teacher signal corresponding to the second information signal is processed to obtain an input signal corresponding to the first information signal. In this case, the quality of the output obtained by the input signal is determined by the adjusted plural kinds of parameter values. For example, when the information signal is an image signal, a plurality of types of parameter values are adjusted to determine the image quality of the image by the input signal. In addition, when the information signal is an audio signal, a plurality of types of parameter values are adjusted to determine the sound quality of the audio by the input signal.

From this input signal, a plurality of first information data located around the target position in the teacher signal is selected, and a class to which the information data of the target position belongs is detected based on the plurality of first information data. Further, based on this input signal, a plurality of second information data located around the target position in the teacher signal is selected.

Then, a plurality of types of parameter values are sequentially adjusted in a plurality of stages, and the class belongs to the class to which the information data of the target position in the teacher signal belongs, the selected plurality of second information data and the information data of the target position in the teacher signal. And for each combination of plural kinds of parameter values, a first normal equation for obtaining coefficient data of the estimation equation is generated, and by solving this equation, coefficient data of the estimation equation for each combination is obtained.

Further, using the coefficient data for each combination, a second normal equation for obtaining coefficient seed data is generated for each class, and the coefficient seed data for each class is obtained by solving this equation.

Here, the coefficient seed data is coefficient data in a generation equation including the above-described plurality of parameters for generating coefficient data used in the estimation equation used when converting from the first information signal to the second information signal. By using this coefficient seed data, it is possible to obtain coefficient data corresponding to a plurality of types of parameter values arbitrarily adjusted by the generation formula. Accordingly, when converting from the first information signal to the second information signal using the estimation equation, the quality of the output obtained by the second information signal can be freely adjusted on the plurality of axes by adjusting a plurality of types of parameter values. have.

Further, the coefficient data generating apparatus according to the present invention is an apparatus for generating coefficient data of an estimation equation used when converting a first information signal composed of a plurality of information data into a second information signal composed of a plurality of information data. Signal processing means for processing the teacher signal corresponding to the second information signal to obtain an input signal corresponding to the first information signal, and parameter adjusting means for adjusting a plurality of types of parameter values for determining the quality of the output obtained by the input signal. And first data selecting means for selecting a plurality of first information data located around the target position in the teacher signal based on the input signal, and a plurality of first information data selected by the first data selecting means. Based on the class detecting means for detecting a class to which the information data of the target position belongs, and based on an input signal, the teacher Second data selecting means for selecting a plurality of second information data located around the target position in the call, a class detected by the class detecting means, a plurality of second information data selected by the second data selecting means, and a teacher signal Normal information generating means for generating a normal equation for obtaining coefficient data of the estimated equation for each combination of class and plural kinds of parameter values using the information data of the point of interest in s, and solving this regular equation for each combination Coefficient data calculating means for obtaining coefficient data.

Further, the coefficient data generation method according to the present invention is a method of generating coefficient data of an estimation equation used when converting a first information signal composed of a plurality of information data into a second information signal composed of a plurality of information data, A first step of processing a teacher signal corresponding to the second information signal to obtain an input signal corresponding to the first information signal, and a second step of adjusting a plurality of types of parameter values for determining the quality of the output obtained by the input signal; And a third step of selecting a plurality of first information data located around the target position in the teacher signal based on the input signal, and based on the plurality of first information data selected in the third step. A fourth step of detecting a class to which the information data of the position belongs, and a plurality of positions located around the target position in the teacher signal based on the input signal; 2 From the fifth step of selecting the information data, the class detected in the fourth step, the plurality of second information data selected in the fifth step and the information data of the target position in the teacher signal, Each combination includes a sixth step of generating a normal equation for obtaining coefficient data of the estimation equation, and a seventh step of solving the normal equation generated in the sixth step to obtain coefficient data for each combination.                 

The information providing medium according to the present invention also provides a computer program for executing each step of the above-described coefficient data generation method.

In the present invention, the teacher signal corresponding to the second information signal is processed to obtain an input signal corresponding to the first information signal. In this case, the quality of the output obtained by the input signal is determined by the adjusted plural kinds of parameter values. For example, when the information signal is an image signal, plural kinds of parameter values are adjusted, and the image quality of the image by the input signal is determined. In addition, when the information signal is a voice signal, a plurality of types of parameter values are adjusted to determine the sound quality of the voice by the input signal.

Based on this input signal, a plurality of first information data located around the target position in the teacher signal is selected, and based on the plurality of first information data, a class to which the information data of the target position belongs is detected. do. Further, based on this input signal, a plurality of second information data located around the point of interest in the teacher signal is selected.

Then, a plurality of types of parameter values are adjusted in a plurality of steps, and a class and a plurality of types are obtained from the class to which the information data of the target position in the teacher signal belongs, the selected plurality of second pixel data and the information data of the target position in the teacher signal. For each combination of parameter values of, a normal equation for obtaining coefficient data of the estimated equation is generated, and by solving this regular equation, coefficient data of the estimated equation for each combination is obtained.

As described above, coefficient data of an estimation equation used when converting the first information signal to the second information signal is generated, but when converting from the first information signal to the second information signal, attention is paid to the second information signal. Coefficient data corresponding to the class to which the information data of the position belongs and the adjusted plural kinds of parameter values are selectively used, and based on the estimation equation, the information data of the position of interest is calculated. Accordingly, when converting from the first information signal to the second information signal using the estimation equation, the quality of the output obtained by the second information signal can be freely adjusted on the plurality of axes by adjusting a plurality of types of parameter values. have.

1 is a block diagram showing the configuration of a TV receiver as an embodiment;

2 illustrates an example of a user interface for adjusting image quality.

3 is an enlarged view of an adjustment screen;

4 is a diagram for explaining a pixel positional relationship between a 525i signal and a 525p signal;

5 is a diagram for explaining a pixel positional relationship between a 525i signal and a 1050i signal;

Fig. 6 is a diagram showing an example of pixel positional relationship and prediction taps of 525i and 525p;

7 is a diagram showing an example of pixel positional relationship and prediction taps of 525i and 525p;

8 shows an example of pixel positional relationships and prediction taps between 525i and 1050i;

9 is a diagram showing an example of pixel positional relationships and prediction taps of 525i and 1050i;

Fig. 10 is a diagram showing an example of pixel positional relationship and spatial class taps of 525i and 525p;

Fig. 11 is a diagram showing an example of pixel positional relationship and spatial class taps of 525i and 525p;                 

Fig. 12 is a diagram showing an example of pixel positional relationship and spatial class taps of 525i and 1050i.

Fig. 13 is a diagram showing an example of pixel positional relationship and spatial class taps of 525i and 1050i;

Fig. 14 is a diagram showing an example of pixel positional relationship and motion class taps of 525i and 525p;

Fig. 15 is a diagram showing an example of pixel positional relationship and motion class taps of 525i and 1050i;

Fig. 16 is a diagram for explaining line double speed processing in the case of outputting a 525p signal.

17 is a diagram illustrating a concept of an example of a method of generating coefficient seed data.

18 is a block diagram showing a configuration example of a coefficient seed data generating device;

19 shows an example of frequency characteristics of a band pass filter.

20 illustrates the concept of another example of a method for generating coefficient seed data.

Fig. 21 is a block diagram showing another configuration example of the coefficient seed data generating device.

22A to 22C are diagrams for explaining a noise adding method.

Fig. 23 is a diagram showing an example of generation of SD signals (parameters r and z).

24 shows an example of an adjustment screen of parameters r and z;

Fig. 25 is a diagram showing an example of generation of SD signals (parameters h, v, z).

Fig. 26 is a diagram showing an example of an adjustment screen of parameters h, v, z.                 

Fig. 27 is a block diagram showing an example of the configuration of an image signal processing apparatus for realizing with software.

Fig. 28 is a flowchart for explaining a processing procedure of an image signal.

Fig. 29 is a flowchart for explaining coefficient seed data generation processing (first).

30 is a flowchart for explaining coefficient seed data generation processing (second).

Fig. 31 is a block diagram showing the configuration of a TV receiver as another embodiment.

32 is a block diagram showing a configuration example of a coefficient seed data generating device;

33 is a flowchart for explaining a processing procedure of an image signal.

34 is a flowchart showing coefficient data generation processing.

35 is a block diagram showing a configuration of an image signal processing device according to another embodiment.

36 is a diagram showing a correspondence relationship between a parameter P value and a function;

EMBODIMENT OF THE INVENTION Hereinafter, embodiment of this invention is described, referring drawings. 1 shows the configuration of a TV receiver 100 as an embodiment. The TV receiver 100 obtains a 525i signal as a standard definition (SD) signal from a broadcast signal, converts the 525i signal into a 525p signal or a 1050i signal as a high definition (HD) signal, and uses the 525p signal or a 1050i signal. Display an image.

Here, the 525i signal means 525 lines and an interlaced image signal, the 525p signal means 525 lines and a progressive (non-interlaced) image signal, and the 1050i signal is 1050 lines and an interlace system Means an image signal.

The TV receiver 100 includes a microcomputer, and has a system controller 101 for controlling the operation of the entire system and a remote control signal receiving circuit 102 for receiving a remote control signal. The remote control signal receiving circuit 102 is connected to the system controller 101, receives the remote control signal RM output by the remote control transmitter 200 according to the user's operation, and transmits an operation signal corresponding to the signal RM to the system controller ( 101).

In addition, the TV receiver 100 is supplied with a reception antenna 105 and a broadcast signal (RF modulated signal) captured by the reception antenna 105, and performs a channel selection process, an intermediate frequency amplification circuit, a detection process, and the like. A tuner 106 for obtaining the above-described SD signal Va (525i signal), an external input terminal 107 for inputting the SD signal Vb (525i signal) from the outside, and selectively outputting any one of these SD signals Va and Vb And a buffer memory 109 for temporarily storing the SD signal outputted from the changeover switch 108.

The SD signal Va output from the tuner 106 is supplied to the fixed terminal a of the selector switch 108, and the SD signal Vb input from the external input terminal 107 is supplied to the fixed terminal b of the selector switch 108. do. The switching operation of the changeover switch 108 is controlled by the system controller 101.

The TV receiver 100 further includes an image signal processing unit 110 for converting an SD signal 525i signal temporarily stored in the buffer memory 109 into an HD signal (525p signal or 1050i signal), and the image signal processing unit 110. Display unit 111 for displaying an image by the HD signal outputted from the control panel), and an OSD (On Screen Display) for generating a display signal SCH for executing display of character graphics or the like on the screen of the display unit 111; Circuit 112 and a synthesizer 113 for synthesizing the display signal SCH with the HD signal output from the above-described image signal processing unit 110 and supplying it to the display unit 111.

The display unit 111 is configured of, for example, a flat panel display such as a cathode-ray tube (CRT) display or a liquid crystal display (LCD). In addition, the generation operation of the display signal SCH in the OSD circuit 112 is controlled by the system controller 101.

The operation of the TV receiver 100 shown in FIG. 1 will be described.

When a mode for executing image display corresponding to the SD signal Va output from the tuner 106 is selected by the user's operation of the remote control transmitter 200, the switching switch 108 is controlled by the control of the system controller 101. It is connected to the a side, and the SD signal Va is output from this changeover switch 108. On the other hand, when a mode for executing the pixel display corresponding to the SD signal Vb input to the external input terminal 107 is selected by the user's operation of the remote control transmitter 200, the changeover switch is controlled by the control of the system controller 101. 108 is connected to the b side, and the SD signal Vb is output from this changeover switch 108.

The SD signal 525i signal output from the changeover switch 108 is written to the buffer memory 109 and temporarily stored. The SD signal temporarily stored in this buffer memory 109 is supplied to the image signal processing unit 110 and converted into an HD signal (525p signal or 1050i signal). That is, the image signal processing unit 110 obtains pixel data (hereinafter referred to as "HD pixel data") constituting the HD signal from pixel data constituting the SD signal (hereinafter referred to as "SD pixel data"). Lose. Here, the selection of the 525p signal or the 1050i signal is performed by the operation of the user's remote control transmitter 200. The HD signal output from the image signal processing unit 110 is supplied to the display unit 111 through the synthesizer 113, and the image by the HD signal is displayed on the screen of the display unit 111.

In addition, although not described above, the user can smoothly adjust the horizontal or vertical resolution of the image displayed on the screen of the display unit 111 steplessly as described above by the operation of the remote control transmitter 200. In the image signal processing unit 110, as described later, HD pixel data is calculated by an estimation equation. As the coefficient data of the estimation equation, the horizontal and vertical resolutions adjusted by the user's operation of the remote control transmitter 200 are determined. Corresponding to h and v to be generated are generated and used by a generation formula including these parameters h and v. Accordingly, the horizontal and vertical resolutions of the image by the HD signal output from the image signal processing unit 110 correspond to the adjusted parameters h and v.

2 shows an example of a user interface for adjusting parameters h and v. At the time of the adjustment, the display screen 111 displays the adjustment screen 115 on which the adjustment positions of the parameters h and v are indicated by an icon 115a of the symbol ☆. Moreover, the remote control transmitter 200 is equipped with the joystick 200a as a user operation means.                 

By operating the joystick 200a, the user can arbitrarily adjust the parameters h and v values for determining the horizontal and vertical resolution by moving the position of the icon 115a on the adjustment screen 115. 3 shows an enlarged portion of the adjustment screen 115. The parameter h value for determining the horizontal resolution is adjusted by moving the icon 115a left and right, while the parameter v value for determining the vertical resolution is adjusted by moving the icon 115a up and down. The user can adjust the parameter h and v values with reference to the adjustment screen 115 displayed on the display 111, and the adjustment can be easily performed.

In addition, the remote control transmitter 200 may be provided with other pointing devices, such as a mouse and a track ball, instead of the joystick 200a. In addition, the parameter h and v values adjusted by the user may be numerically displayed on the adjustment screen 115.

Next, the image signal processing unit 110 will be described in detail. The image signal processing unit 110 stores data of a plurality of SD pixels located around the pixel of interest in the HD signal (1050i signal or 525p signal) from the SD signal 525i signal stored in the buffer memory 109. Has first to third tap selector circuits 121 to 123 for selectively extracting and outputting the?

The first tap selection circuit 121 selectively extracts data of an SD pixel (called "prediction tap") used for prediction. The second tap selection circuit 122 selectively extracts data of SD pixels (referred to as "spatial class taps") used for class classification corresponding to the level distribution pattern of the SD pixel data. The third tap selection circuit 123 selectively extracts data of SD pixels (called "motion class taps") used for class classification corresponding to movement. When the spatial class is determined using SD pixel data belonging to a plurality of fields, the motion class is also included in this spatial class.

Fig. 4 shows the pixel positional relationship of the odd (o) field of an arbitrary frame F of the 525i signal and the 525p signal. The large dot is the pixel of the 525i signal, and the small dot is the pixel of the 525p signal outputted. In the even field, the lines of the 525i signal are shifted by 0.5 lines spatially. As can be seen from Fig. 4, the pixel data of the 525p signal includes line data L1 at the same position as the line of the 525i signal and line data L2 at the intermediate position of the upper and lower lines of the 525i signal. In addition, the number of pixels of each line of the 525p signal is twice the hydrogen of each line of the 525i signal.

Fig. 5 shows the pixel positional relationship of an arbitrary frame F of the 525i signal and the 1050i signal. The pixel positions of the odd (o) field are indicated by solid lines, and the pixel positions of the even (e) field are shown by broken lines. The large dot is the pixel of the 525i signal, and the small dot is the pixel of the 1050i signal outputted. As can be seen from Fig. 5, the pixel data of the 1050i signal includes line data L1, L1 'at a position close to the line of the 525i signal, and line data L2, L2' at a position far from the line of the 525i signal. Here, L1 and L2 are line data of odd field, and L1 'and L2' are line data of even field. The number of pixels in each line of the 1050i signal is twice the number of pixels in each line of the 525i signal.

6 and 7 show specific examples of prediction taps (SD pixels) selected by the first tap selection circuit 121 when converting a 525i signal to a 525p signal. 6 and 7 show pixel positional relationships in the vertical direction of odd (o) and even (e) fields of frames F-1, F, and F + 1 that are continuous in time.

As shown in Fig. 6, the prediction tap at the time of predicting the line data L1 and L2 of the field F / o is included in the next field F / e, and the pixels of the 525p signal to be created (pixels at the attention position) SD pixels T1, T2, and T3, which are spatially close to each other, and SD pixels T4, T5, and T6, which are spatially near to the pixels of the 525p signal included in the field F / o and must be created, and the preceding field F-. SD pixels T7, T8, and T9 that are contained in 1 / e and are spatially near the pixel of the 525p signal to be created, and the pixel of the 525p signal to be included in the earlier field F-1 / o. The SD pixel T10 is spatially in the vicinity of.

As shown in FIG. 7, the prediction tap at the time of predicting the line data L1, L2 of the field F / e is included in the next field F + 1 / o, and is spatially near the pixel of the 525p signal to be created. SD pixels T1, T2, and T3, included in field F / e, and included in SD pixels T4, T5, and T6, which are spatially close to the pixels of the 525p signal to be created, and in the previous field F / o. SD pixels T7, T8, and T9, which are spatially close to the pixels of the 525p signal to be created, and spatially near the pixels of the 525p signal, which are contained in the earlier field F-1 / e. SD pixel T10.

When predicting the line data L1, the SD pixel T9 may not be selected as the prediction tap, while when predicting the line data L2, the SD pixel T4 may not be selected as the prediction tap.

8 and 9 show specific examples of prediction taps (SD pixels) selected by the first tap selection circuit 121 when the 525i signal is converted into a 1050i signal. 8 and 9 show pixel positional relationships in the vertical direction of odd (o) and even (e) fields of frames F-1, F, and F + 1 that are continuous in time.

As shown in Fig. 8, the prediction tap at the time of predicting the line data L1, L2 of the field F / o is included in the next field F / e, and the pixels of the 1050i signal (pixels of note position) to be created are SD pixels T1, T2, which are spatially close to each other, SD pixels T3, T4, T5, T6, which are spatially close to the pixels of the 1050i signal contained in the field F / o to be created, and the preceding field F-. SD pixels T7 and T8 that are included in 1 / e and are spatially close to the pixels of the 1050i signal to be created.

As shown in Fig. 9, the prediction tap when predicting the line data L1 'and L2' of the field F / e is included in the next field F + 1 / o and spatially with respect to the pixel of the 1050i signal to be created. SD pixels T1, T2, which are in the near position, SD pixels T3, T4, T5, and T6, which are included in the field F / e and spatially in the vicinity of the pixel of the 1050i signal to be created, and in the previous field F / o. SD pixels T7 and T8 that are included and are spatially close to the pixels of the 1050i signal to be created.

When predicting the line data L1 ', the SD pixel T6 may not be selected as the prediction tap, while when predicting the line data L2', the SD pixel T3 may not be selected as the prediction tap.

6 to 9, one or a plurality of SD pixels in the horizontal direction may be selected as prediction taps in addition to the SD pixels at the same positions in the plurality of fields.

10 and 11 show specific examples of the space class taps (SD pixels) selected by the second tap selector 122 when converting from a 525i signal to a 525p signal. 10 and 11 show pixel positional relationships in the vertical direction of odd (o) and even (e) fields of frames F-1, F, and F + 1 that are temporally continuous.

As shown in Fig. 10, the spatial class tap when predicting the line data L1 and L2 of the field F / o is included in the next field F / e, and the pixel of the 525p signal to be created (the pixel at the note position) is created. SD pixels T1, T2, which are spatially close to each other, SD pixels T3, T4, T5, which are included in the field F / o and spatially close to the pixels of the 525p signal to be created, and the preceding field F-1. SD pixels T6 and T7 that are included in / e and are spatially close to the pixels of the 525p signal to be created.

As shown in Fig. 11, the spatial class tap when predicting the line data L1 and L2 of the field F / e is included in the next field F + 1 / o and spatially near the pixel of the 525p signal to be created. Included in the SD pixels T1, T2, which are positions, and in the fields F / e, and included in the SD pixels T3, T4, T5, which are spatially close to the pixels of the 525p signal to be created, and in the previous fields F / o, SD pixels T6 and T7 that are spatially close to the pixels of the 525p signal to be created.

In addition, the SD pixel T7 may not be selected as the spatial class tap when predicting the line data L1, and the SD pixel T6 may not be selected as the spatial class tap when predicting the line data L2.                 

12 and 13 show specific examples of the space class taps (SD pixels) selected by the second tap selection circuit 122 when the 525i signal is converted into a 1050i signal. 12 and 13 show pixel positional relationships in the vertical direction of odd (o) and even (e) fields of frames F-1, F, and F + 1 that are continuous in time.

As shown in Fig. 12, the spatial class taps for predicting the line data L1 and L2 of the field F / o are included in the field F / o, and the pixels of the 1050i signal (pixels of note position) to be created are shown. SD pixels T1, T2, and T3, which are spatially close to each other, and SD pixels T4, T5, T6, and T7, which are included in the preceding field F-1 / e and are spatially close to the pixels of the 1050i signal to be created. .

As shown in Fig. 13, the spatial class taps when predicting the line data L1 'and L2' of the field F / e are included in the field F / e and spatially near the pixel of the 1050i signal to be created. SD pixels T1, T2, and T3, and SD pixels T4, T5, T6, and T7, which are included in the previous field F / o and are spatially close to the pixels of the 1050i signal to be created.

Note that when predicting the line data L1 ', the SD pixel T7 may not be selected as the spatial class tap, while when predicting the line data L2', the SD pixel T4 may not be selected as the spatial class tap.

10 to 13, one or a plurality of SD pixels in the horizontal direction may be selected as the space class taps in addition to the SD pixels at the same positions in the plurality of fields.                 

FIG. 14 shows a specific example of a motion class tap (SD pixel) selected by the third tap selector circuit 123 when the 525i signal is converted into a 525p signal. Fig. 14 shows the pixel positional relationship in the vertical direction of odd (o) and even (e) fields of frames F-1 and F, which are continuous in time. As shown in Fig. 14, the motion class tap when predicting the line data L1 and L2 of the field F / o is included in the next field F / e, and is applied to the pixel of the 525p signal (pixel at the attention position) to be created. SD pixels n2, n4, n6, which are spatially close to each other, and SD pixels n1, n3, n5, which are included in the field F / o and are spatially close to the pixels of the 525p signal to be created, and the preceding field F S / D pixels m2, m4, and m6, which are spatially located relative to the pixel of the 525p signal to be written and contained in -1 / e, and in the earlier field F-1 / o, which are to be written 525p The SD pixels m1, m3, and m5 are spatially near the pixels of the signal. The vertical position of each of the SD pixels n1 to n6 coincides with the vertical position of each of the SD pixels m1 to m6.

FIG. 15 shows a specific example of a motion class tap (SD pixel) selected by the third tap selection circuit 123 when the 525i signal is converted into a 1050i signal. Fig. 15 shows the pixel positional relationship in the vertical direction of odd (o) and even (e) fields of frames F-1 and F that are continuous in time. As shown in Fig. 15, the motion class tap when predicting the line data L1 and L2 of the field F / o is included in the next field F / e and is located in the spatial vicinity of the pixel of the 1050i signal to be created. Included in the SD pixels n2, n4, n6 and the SD pixels n1, n3, n5, which are included in the field F / o, and are spatially close to the pixels of the 1050i signal to be created, and in the preceding field F-1 / e. S / D pixels m2, m4, and m6 that are spatially close to the pixels of the 1050i signal to be created, and the pixels of the 1050i signal to be created and included in the earlier field F-1 / o. These are SD pixels m1, m3, and m5 in the vicinity of each other. The vertical position of each of the SD pixels n1 to n6 coincides with the vertical position of each of the SD pixels m1 to m6.

Returning to FIG. 1, the image signal processing unit 110 detects the level distribution pattern of the data (SD pixel data) of the spatial class taps selectively extracted by the second tap selection circuit 122, and applies the level distribution pattern to the level distribution pattern. The space class detection circuit 124 detects a space class and outputs the class information.

In the space class detection circuit 124, for example, an operation of compressing each SD pixel data from 8-bit data to 2-bit data is executed. The space class detection circuit 124 then outputs compressed data corresponding to each SD pixel data as class information of the space class. In this embodiment, data compression is performed by adaptive dynamic range coding (ADRC). As the information compression means, in addition to ADRC, DPCM (prediction coding), VQ (vector quantization), or the like may be used.

Originally, ADRC is an adaptive quantization method developed for high performance coding for video tape recorders (VTRs), but is suitable for use in the above-mentioned data compression because it can efficiently represent local patterns of signal levels with short word lengths. . When using ADRC, set the maximum value of the data (SD pixel data) of the spatial class tap to MAX, the minimum value to MIN, the dynamic range of the spatial class tap to DR (= MAX-MIN + 1), and the number of requantization bits to P. In this case, the requantization code qi as compressed data is obtained by the calculation of equation (1) for each SD pixel data ki as data of the spatial class tap. However, in Equation 1, [] means truncation processing. I = 1 to Na when there are Na pieces of SD pixel data as data of the spatial class tap.

In addition, the image signal processing unit 110 detects a motion class mainly for expressing the degree of motion from the data (SD pixel data) of the motion class taps selectively extracted by the third tap selection circuit 123, and the class It has a motion class detection circuit 125 for outputting information.

In this motion class detecting circuit 125, the interframe difference is calculated from the data (SD pixel data) mi and ni of the motion class tap selectively extracted by the third tap selecting circuit 123, and the absolute value of the difference is calculated. Threshold processing is performed on the average value of to detect a motion class that is an indicator of motion. That is, in the motion class detection circuit 125, the average value AV of the absolute values of the difference is calculated by the equation (2). In the third tap selection circuit 123, for example, when twelve SD pixel data m1 to m6 and n1 to n6 are extracted as described above, Nb in Equation (2) is six.

In the motion class detection circuit 125, the average value AV calculated as described above is compared with one or a plurality of threshold values to obtain the motion information MV of the motion class. For example, three threshold values th1, th2, and th3 (th1 <th2 <th3) are prepared, and when four motion classes are detected, MV = 0 when AV≤th1 and MV when th1 <AV≤th2. = 1, MV = 2 when th2 <AV <th3, and MV = 3 when th3 <AV.

In addition, the image signal processing unit 110 is based on the requantization code qi as class information of the spatial class output from the spatial class detection circuit 124 and the motion information MV of the motion class output from the motion class detection circuit 125. And a class synthesizing circuit 126 for obtaining class code CL indicating the class of the pixel data of the HD signal (525p signal or 1050i signal) to be created, that is, the pixel data of the target position.

In this class synthesizing circuit 126, the calculation of the class code CL is performed by the expression (3). In Equation 3, Na represents the number of data (SD pixel data) of the spatial class tap, and P represents the number of requantization bits in ADRC.

In addition, the image signal processing unit 110 has registers 130 to 131 and a coefficient memory 134. The line sequential conversion circuit 129 described later needs to switch the operation when outputting a 525p signal and outputting a 1050i signal. The register 130 stores operation specification information for specifying the operation by the line sequential conversion circuit 129. The line sequential converting circuit 129 operates according to the operation designation information supplied from the register 130.

The register 131 stores the tap position information of the prediction tap selected by the first tap selection circuit 121. The first tap selection circuit 121 selects the prediction tap according to the tap position information supplied from the register 131. The tap position information is, for example, given a number to a plurality of SD pixels that may be selected, and designates the number of the selected SD pixels. The same applies to the following tap position information.

The register 132 stores the tap position information of the space class tap selected by the second tap selector 122. The second tap selection circuit 122 selects the space class taps according to the tap position information supplied from the register 132.

Here, the register 132 stores the tap position information A when the movement is relatively small and the tap position information B when the movement is relatively large. Which of these tap positional information A, B is supplied to the second tap selector 122 is selected by the motion information MV of the motion class output from the motion class detection circuit 125.

That is, when MV = 0 or MV = 1 because there is no movement or the movement is small, tap position information A is supplied to the second tap selection circuit 122, and the second tap selection circuit 122 is selected. The spatial class taps cover a plurality of fields as shown in Figs. In addition, when the movement is relatively large and MV = 2 or MV = 3, the tap position information B is supplied to the second tap selection circuit 122, and the spatial class tap selected by the second tap selection circuit 122 is shown. If not, only the SD pixels in the same field as the pixels to be created are included.

The tap position information when the movement is relatively small and the tap position information when the movement is relatively large are also stored in the register 131 described above, and the tap position information supplied to the first tap selection circuit 122 is used as the motion class. The motion information MV of the motion class output from the detection circuit 125 may be selected.

The register 133 stores tap position information of the motion class tap selected by the third tap selection circuit 123. The third tap selection circuit 123 selects the motion class tap according to the tap position information supplied from the register 133.

The coefficient memory 134 stores coefficient data of an estimation equation used in the estimation prediction calculating circuit 127 described later for each class. This coefficient data is information for converting a 525i signal as an SD signal into a 525p signal or a 1050i signal as an HD signal. The class code CL output from the above-described class synthesizing circuit 126 is supplied as read address information to the coefficient memory 134, and coefficient data corresponding to the class code CL is read out from the coefficient memory 134 to estimate estimation operation. Supplied to the circuit 127.

In addition, the image signal processing unit 110 has an information memory bank 135. In this information memory bank 135, operation designation information for storing in the register 130 and tap position information for storing in the registers 131 to 133 are stored in advance.

Here, as the operation designation information for storing in the register 130, the information memory bank 135 includes first operation designation information for operating the line sequential conversion circuit 129 to output a 525p signal, and a line sequential conversion circuit ( Second operation designation information for operating 129 to output the 1050i signal is stored in advance.

By operating the remote control transmitter 200, the user can select a first conversion method for outputting a 525p signal as an HD signal or a second conversion method for outputting a 1050i signal as an HD signal. The system controller 101 is supplied with selection information of the conversion method to the information memory bank 135, and the register 130 is supplied from the information memory bank 135 to the first operation designation information or the second operation according to the selection information. The assignment information is loaded.

The information memory bank 135 further includes first tap position information and second transform method 1050i corresponding to the first conversion method 525p as tap position information of the predicted tap for storing in the register 131. Corresponding second tap position information is stored in advance. In this information memory bank 135, the first tap position information or the second tap position information is loaded into the register 131 according to the above selection information of the conversion method.

In addition, the information memory bank 135 includes the tap position information of the space class taps to be stored in the register 132 to the first tap position information and the second converting method 1050i corresponding to the first converting method 525p. Corresponding second tap position information is stored in advance. The first and second tap position information each consist of tap position information when the movement is relatively small and tap position information when the movement is relatively large. In this information memory bank 135, the first tap position information or the second tap position information is loaded in the register 132 in accordance with the selection information of the above-described conversion method.                 

The information memory bank 135 also includes tap position information of the motion class taps to be stored in the register 133, and corresponds to the first tap position information and the second converting method 1050i corresponding to the first converting method 525p. Corresponding second tap position information is stored in advance. In this information memory bank 135, the first tap position information or the second tap position information is loaded into the register 133 in accordance with the above selection information of the conversion method.

In addition, the coefficient memory data of each class corresponding to each of the first and second conversion methods is stored in the information memory bank 135 in advance. This coefficient seed data is coefficient data of a generation formula for generating coefficient data for storing in the coefficient memory 134 described above.

In the estimation prediction calculating circuit 127 described later, the HD pixel data y to be generated by the estimation formula of Equation 4 is obtained from the data (SD pixel data) xi of the prediction tap and the coefficient data Wi read out from the coefficient memory 134. Is calculated. When the number of prediction taps selected by the first tap selection circuit 121 is ten as shown in FIGS. 4 and 7, n in Equation 4 becomes 10.

The coefficient data Wi (i = 1 to n) of this estimation equation is generated by a generation equation including parameters h and v, as shown in equation (5). In the information memory bank 135, each coefficient data, the coefficient seed how the data w ₁₀ ~w _n9 conversion of the production equation, are also stored for each class. The generation method of this coefficient seed data is mentioned later.

In addition, the image signal processing unit 110 uses the coefficient type data of each class and the parameter h and v values, and coefficient data Wi (i = 1) of the estimation equation corresponding to the parameter h and v values for each class according to equation (5). And coefficient generation circuit 136 for generating ˜n). In the coefficient generation circuit 136, the coefficient type data of each class corresponding to the first conversion method or the second conversion method is loaded in the information memory bank 135 in accordance with the above selection information of the conversion method. In addition, the coefficient generation circuit 136 is supplied with the parameter h and v values from the system controller 101.

The coefficient data Wi (i = 1 to n) of each class generated by this coefficient generation circuit 136 is stored in the coefficient memory 134 described above. The generation of coefficient data Wi of each class in this coefficient generation circuit 136 is performed in each vertical blanking period, for example. Accordingly, even if the parameter h and v values are changed by the operation of the user's remote control transmitter 200, the coefficient data Wi of each class stored in the coefficient memory 134 is immediately changed to correspond to the parameter h and v values. The adjustment of the resolution by the user can be performed smoothly.

The image signal processing unit 110 further includes a normalization coefficient generation circuit for calculating the normalization coefficient S corresponding to the coefficient data Wi (i = 1 to n) of each class generated by the coefficient generation circuit 136 by equation (6). 137 and a normalization coefficient memory 138 for storing the normalization coefficient S generated here for each class. The class code CL output from the above-described class synthesizing circuit 126 is supplied as read address information to the normalization coefficient memory 138, and the normalization coefficient S corresponding to the class code CL is read out from the normalization coefficient memory 138. And supplied to the normalization arithmetic circuit 128 described later.

In addition, the image signal processing unit 110 should create from the data (SD pixel data) xi of the prediction tap selectively extracted by the first tap selection circuit 121 and the coefficient data Wi read out from the coefficient memory 134. There is an estimation prediction calculating circuit 127 that calculates data of pixel data (pixel data at the main position) of the HD signal.

In the estimation prediction calculating circuit 127, when outputting the 525p signal, as shown in FIG. 4, the line data L1 at the same position as the line of the 525i signal in the odd (o) field and the even (e) field. It is necessary to generate line data L2 at the intermediate position of the upper and lower lines of the 525i signal, and double the number of pixels of each line. When the 1050i signal is output by the estimation prediction calculating circuit 127, as shown in FIG. 5 described above, a line at a position close to the line of the 525i signal in the odd (o) field and the even (e) field. It is necessary to generate the line data L2, L2 'at a position far from the lines of the data L1, L1' and the 525i signal, and double the number of pixels in each line.

Therefore, the estimation prediction calculating circuit 127 simultaneously generates data of four pixels constituting the HD signal. For example, the data of four pixels are generated simultaneously by using estimation equations in which coefficient data are different, respectively, and coefficient data of each estimation equation is supplied from the coefficient memory 134. In the estimation prediction calculating circuit 127, the HD pixel to be created by the above-described estimation formula (4) from the data of the prediction tap (SD pixel data) xi and the coefficient data Wi read out from the coefficient memory 134. The data y is calculated.

In addition, the image signal processing unit 110 reads from the normalization coefficient memory 138 the HD pixel data y constituting the line data L1 and L2 (L1 ', L2') output from the estimation prediction calculating circuit 127. It has a normalization arithmetic circuit 128 which divides and normalizes to the normalization coefficient S corresponding to the coefficient data Wi (i = 1 to n) used for each generation. Although not described above, the coefficient generation circuit 136 obtains the coefficient data of the estimation equation from the coefficient seed data in the generation equation. The generated coefficient data includes a rounding error, and the sum of the coefficient data Wi (i = 1 to n) is It is not guaranteed to be 1.0. Therefore, the HD pixel data y calculated by the estimation prediction calculating circuit 127 is level shifted due to the rounding error. As described above, the change can be eliminated by normalizing by the normalization calculating circuit 128.                 

In addition, the image signal processing unit 110 executes the line double speed process of doubling the horizontal period, so that the line data L1, L2 (L1 ', which is supplied from the estimation prediction calculating circuit 127 through the normalization calculating circuit 128). L2 ') has a line sequential converting circuit 129 for line ordering.

Fig. 16 shows the line double speed processing in the case of outputting a 525p signal using an analog waveform. As described above, the line data L1 and L2 are generated by the estimation prediction calculating circuit 127. In line data L1, a1, a2, a3,. Lines are included, and the line data L2 is sequentially b1, b2, b3,... Line of. The line sequential converting circuit 129 compresses the data of each line in half in the time axis direction, and alternately selects the compressed data to select the line sequential outputs a0, b0, a1, b1,... To form.

When the 1050i signal is output, the line sequential converter 129 generates the line sequential output so as to satisfy the interlace relationship in the odd field and the even field. Therefore, the line sequential conversion circuit 129 needs to switch its operation when outputting the 525p signal and when outputting the 1050i signal. The operation designation information is supplied from the register 130 as described above.

Next, the operation of the image signal processing unit 110 will be described.

From the SD signal 525i signal stored in the buffer memory 109, the data (SD pixel data) of the spatial class taps are selectively extracted by the second tap selection circuit 122. In this case, the second tap selection circuit 122 taps on the basis of the tap position information corresponding to the motion method detected by the conversion method selected by the user and the motion class detection circuit 125 supplied from the register 132. Selection is executed.

The data (SD pixel data) of the spatial class tap selectively extracted by the second tap selection circuit 122 is supplied to the spatial class detection circuit 124. In this spatial class detection circuit 124, ADRC processing is performed on each SD pixel data as data of a spatial class tap to obtain a requantization code qi as class information of a spatial class (mainly class classification for waveform representation in space). (See Equation 1).

The third tap selector 123 selectively extracts the motion class tap data (SD pixel data) from the SD signal 525i signal stored in the buffer memory 109. In this case, the third tap selection circuit 123 performs tap selection based on tap position information corresponding to the conversion method selected by the user, supplied from the register 133.

The data (SD pixel data) of the motion class tap selectively extracted by the third tap selection circuit 123 is supplied to the motion class detection circuit 125. In this motion class detection circuit 125, motion information MV of a motion class (mainly class classification for expressing the degree of motion) is obtained from each SD pixel data as data of the motion class tap.

This motion information MV and the requantization code qi described above are supplied to the class synthesizing circuit 126. In this class synthesizing circuit 126, from these motion information MVs and the requantization code qi, a class code CL indicating a class to which the pixel data (pixel data at the note position) of the HD signal (525p signal or 1050i signal) to be generated belongs. Obtained (see Equation 3). This class code CL is supplied to the coefficient memory 134 and the normalization coefficient memory 138 as read address information.

In the coefficient memory 134, coefficient data Wi (i = 1 to n) of the estimation formula of each class corresponding to the parameters h and v values adjusted by the user in each vertical blanking period is, for example, the coefficient generating circuit 136. Is generated and stored. In the normalization coefficient memory 138, the normalization coefficient S corresponding to the coefficient data Wi (i = 1 to n) of each class generated by the coefficient generation circuit 136 is generated by the normalization coefficient generation circuit 137 as described above. Created and stored.

By supplying the class code CL as read address information as described above to the coefficient memory 134, the coefficient data Wi corresponding to the class code CL is read out from this coefficient memory 134 and supplied to the estimation prediction calculating circuit 127. . Further, data of the prediction tap (SD pixel data) is selectively extracted by the first tap selector 12 from the SD signal 525i signal stored in the buffer memory 109. In this case, the first tap selection circuit 121 selects the tap based on the tap position information corresponding to the conversion method selected by the user, supplied from the register 131. The data (SD pixel data) xi of the prediction tap selectively extracted by the first tap selection circuit 121 is supplied to the estimation prediction calculating circuit 127.

In the estimation prediction calculating circuit 127, from the data (SD pixel data) xi of the prediction tap and the coefficient data Wi read out from the coefficient memory 134, pixel data of the HD signal to be created, that is, pixel data of the target position (HD) Pixel data) y is calculated (see Equation 4). In this case, data of four pixels constituting the HD signal are simultaneously generated.

Accordingly, when the first conversion method for outputting the 525p signal is selected, the intermediate position of the line data L1 at the same position as the line of the 525i signal and the up and down line of the 525i signal in the odd (o) field and the even (e) field is selected. Line data L2 is generated (see Fig. 4). Further, when the second conversion method for outputting the 1050i signal is selected, it is far from the line of the line data L1, L1 'and the 525i signal at positions close to the line of the 525i signal in the odd (o) field and the even (e) field. Line data L2, L2 'at the position is generated (see Fig. 5).

In this way, the line data L1 and L2 (L1 ', L2') generated by the estimation prediction calculating circuit 127 are supplied to the normalization calculating circuit 128. As described above in the normalization coefficient memory 138, the class code CL is supplied as read address information, so that the normalization coefficient S corresponding to the class code CL is output from the normalization coefficient memory 138, i.e., the estimation prediction calculating circuit 127. The normalization coefficient S corresponding to the coefficient data Wi (i = 1 to n) used in the generation of the respective HD pixel data y constituting the line data L1 and L2 (L1 ', L2') to be read is read and normalized. Is supplied. In the normalization arithmetic circuit 128, each HD pixel data y constituting the line data L1, L2 (L1 ', L2') output from the estimation prediction arithmetic circuit 127 is divided by a corresponding normalization coefficient S, and normalized. . Accordingly, the level variation of the pixel data at the target position due to the rounding error when the coefficient data used in the estimation equation (see Equation 4) is obtained by the generation equation (see Equation 5) using the coefficient seed data is eliminated. .

In this way, the line data L1 and L2 (L1 ', L2') normalized by the normalization arithmetic circuit 128 are supplied to the line sequential conversion circuit 129. In this line sequential conversion circuit 129, the line data L1 and L2 (L1 ', L2') are line-sequentially generated to generate an HD signal. In this case, the line sequential conversion circuit 129 operates according to the operation instruction information corresponding to the conversion method selected by the user, supplied from the register 130. Therefore, when the first conversion method 525p is selected by the user, the sequential conversion circuit 129 outputs a 525p signal. On the other hand, when the second conversion method 1050i is selected by the user, the sequential conversion circuit 129 outputs a 1050i signal.

As described above, the coefficient data Wi (i = 1 to n) of the estimated equation corresponding to the parameter h and v values for each class using the coefficient seed data loaded from the information memory bank 135 in the coefficient generation circuit 136. Is generated and stored in the coefficient memory 134. The HD pixel data y is calculated by the estimation prediction calculating circuit 127 using coefficient data Wi (i = 1 to n) read out in correspondence with the class code CL in the coefficient memory 134.

In this manner, in the image signal processing unit 110, coefficient data Wi (i = 1 to n) of an estimated equation corresponding to the adjusted plural types of parameters h and v values is used, and HD image data y is calculated. Therefore, the user can freely adjust the image quality of the image by the HD signal on the axis of horizontal resolution and vertical resolution by adjusting the parameters h and v values. In addition, the user can smoothly adjust the horizontal and vertical image quality of the image by the HD signal steplessly by adjusting the parameters h and v values. In this case, coefficient data of each class corresponding to the adjusted parameter h and v values is generated and used by the coefficient generation circuit 136 each time, and a memory for storing a large amount of coefficient data is required. I never do that.

As described above, the coefficient seed data is stored for each conversion method and for each class in the information memory bank 135. This coefficient seed data is generated by learning in advance.

First, an example of this generation method is described. The group is an example to obtain the coefficient data in the coefficient seed data w ₁₀ ~w _n9 the generation equation of the equation (5).

Here, ti (i = 0 to 9) is defined as in Equation 7 for the following description.

Using this equation (7), equation (5) can be rewritten as shown in equation (8).

Finally, the unknown coefficient w _xy is obtained by learning. That is, a coefficient value for minimizing the squared error is obtained using a plurality of SD pixel data and HD pixel data for each conversion method and for each class. The so-called least-squares solution. If the number of learning is m and the residual in the k (1≤k≤m) th learning data is e _k and the sum of squared errors is E, E is expressed by Equation 9 using Equations 4 and 5 . Here, x _ik denotes k-th pixel data at the i-th prediction tap position of the SD image, and y _k denotes pixel data of the k-th HD image corresponding thereto.

In the least squares solution, w _xy is obtained, where the partial derivative by w _xy in Expression (9) becomes zero. This is represented by equation (10).

Hereinafter, if X _ipjq and Y _ip are defined as in Equations 11 and 12, Equation 10 can be rewritten as Equation 13 using a matrix.

This equation is commonly called a regular equation. This normal equation is solved for w _xy using Gauss-Jordan cancellation and the like and coefficient coefficient data is calculated.

Fig. 17 shows the concept of the above-described generation method of coefficient seed data. Generate a plurality of SD signals from the HD signal. For example, a total of 81 types of SD signals are generated by varying the parameters h and v that change the horizontal and vertical bands of the filter used when generating the SD signal from the HD signal in nine steps. Learning is performed between the plurality of SD signals and the HD signals generated in this way to generate coefficient seed data.

18 shows the configuration of the coefficient seed data generating device 150 for generating coefficient seed data in the above-described concept.                 

The coefficient seed data generating device 150 performs an input terminal 151 to which an HD signal (525p signal / 1050i signal) as a teacher signal is input, and performs horizontal and vertical thinning processing on the HD signal to perform SD as an input signal. It has an SD signal generation circuit 152 for obtaining a signal.

The SD signal generation circuit 152 is supplied with a conversion method selection signal as a control signal. When the first conversion method (obtaining a 525p signal from the 525i signal in the image signal processing unit 110 in Fig. 1) is selected, the SD signal generation circuit 152 performs thinning processing on the 525p signal to generate an SD signal. (See Figure 4). On the other hand, when the second conversion method (the 1050i signal is obtained from the 525i signal in the image signal processing unit 110 of FIG. 1) is selected, the SD signal generation circuit 152 performs thinning processing on the 1050i signal so that the SD signal is generated. Generated (see FIG. 5).

In addition, the parameters h and v are supplied to the SD signal generation circuit 152 as a control signal. Corresponding to the parameters h and v, the horizontal band and vertical band of the filter used when generating the SD signal from the HD signal are varied. Here, some examples are shown about the detail of a filter.

For example, the case where a filter is comprised by the band filter which limits a horizontal band, and the band filter which limits a vertical band is considered. In this case, as shown in FIG. 19, by designing a frequency characteristic corresponding to the stepped value of the parameter h or v and performing inverse Fourier transform, the one-dimensional filter having the frequency characteristic corresponding to the stepped value of the parameter h or v Can be obtained.                 

Further, for example, a case may be considered in which the filter is composed of a one-dimensional Gaussian filter that limits the horizontal band and a one-dimensional Gaussian filter that limits the vertical band. This one-dimensional Gaussian filter is represented by equation (14). In this case, a one-dimensional Gaussian filter having a frequency characteristic corresponding to the stepped value of the parameter h or v can be obtained by changing the standard deviation σ stepwise corresponding to the stepped value of the parameter h or v.

Further, for example, a case may be considered in which the filter is composed of two-dimensional filters F (h, v) in which horizontal and vertical frequency characteristics are determined at both sides of the parameters h and v. This two-dimensional filter generation method is similar to the one-dimensional filter described above, by designing a two-dimensional frequency characteristics corresponding to the step values of the parameters h, v and performing a two-dimensional inverse Fourier transform, thereby the step values of the parameters h, v A two-dimensional filter having a two-dimensional frequency characteristic corresponding to can be obtained.

In addition, the coefficient seed data generating device 150 includes a plurality of SDs positioned around the target position in the HD signal (1050i signal or 525p signal) from the SD signal (525i signal) output from the SD signal generation circuit 152. First to third tap selection circuits 153 to 155 for selectively extracting and outputting data of pixels.

These first to third tap select circuits 153 to 155 are configured in the same manner as the first to third tap select circuits 121 to 123 of the image signal processing unit 110 described above. The taps selected by the first to third tap select circuits 153 to 155 are designated by tap position information from the tap select control unit 156.

The tap selection control circuit 156 is supplied with a conversion method selection signal as a control signal. In the case where the first conversion method is selected and the second conversion method is selected, the tap position information supplied to the first to third tap selection circuits 153 to 155 is different. The tap selection control circuit 156 is supplied with the motion information MV of the motion class output from the motion class detection circuit 158 described later. Accordingly, the tap position information supplied to the second tap selection circuit 154 may vary depending on whether the movement is large or small.

In addition, the coefficient seed data generating device 150 detects a level distribution pattern of the data (SD pixel data) of the spatial class tap selectively extracted by the second tap selection circuit 154, and based on the level distribution pattern The space class detection circuit 157 is provided for detecting the space class and outputting the class information. This space class detection circuit 157 is configured in the same manner as the space class detection circuit 124 of the image signal processing unit 110 described above. In this space class detection circuit 157, the requantization code qi for each SD pixel data as the data of the space class tap is output as class information indicating the space class.

In addition, the coefficient seed data generating device 150 detects a motion class mainly for expressing the degree of motion, from the data of the motion class tap selectively extracted by the third tap selection circuit 155 (SD pixel data). And a motion class detection circuit 158 for outputting this motion information MV. This motion class detection circuit 158 is configured similarly to the motion class detection circuit 125 of the image signal processing unit 110 described above. In this motion class detection circuit 158, the interframe difference is calculated from the data (SD pixel data) of the motion class tap selectively extracted by the third tap selection circuit 155, and further, the average value of the absolute values of the difference. Threshold processing is performed for to detect a motion class that is an indicator of motion.

The coefficient seed data generating device 150 further includes requantization code qi as class information of the spatial class output from the spatial class detection circuit 157 and motion information MV of the motion class output from the motion class detection circuit 158. On the basis of this, a class synthesizing circuit 159 for obtaining a class code CL indicating a class to which the pixel data of the target position in the HD signal (525p signal or 1050i signal) belongs. This class synthesizing circuit 159 is also configured similarly to the class synthesizing circuit 126 of the image signal processing unit 110 described above.

In addition, the coefficient seed data generating device 150 selects each HD pixel data y as pixel data of the target position obtained from the HD signal supplied to the input terminal 151 and the first tap selection corresponding to each of the HD pixel data y, respectively. Coefficient seed data for each class from data (SD pixel data) xi of the prediction tap selectively extracted by the circuit 153 and class code CL output from the class synthesizing circuit 159 corresponding to each HD pixel data y, respectively. A regular equation generator 160 generates a normal equation (see Equation 13) for obtaining w _{10 to} w _n9 .

In this case, the training data is generated from a combination of one HD pixel data y and n prediction tap pixel data corresponding thereto, but the parameters h and v to the SD signal generation circuit 152 are sequentially changed so that horizontal and vertical bands are provided. A plurality of SD signals changed in steps are sequentially generated, and accordingly, the normal equation generator 160 generates a normal equation in which a lot of learning data is registered.

Here, the coefficient seed data calculated by learning between the HD signal and the SD signal generated by applying a narrow band filter from the HD signal is for obtaining an HD signal with high resolution. On the contrary, coefficient seed data calculated by learning between an HD signal and an SD signal generated by operating a wide band filter from the HD signal is used to obtain an HD signal having a low resolution. As described above, by generating a plurality of SD signals sequentially and registering the learning data, coefficient seed data for obtaining an HD signal of continuous resolution can be obtained.

Although not shown, the SD pixel data xi supplied from the first tap selection circuit 153 to the regular equation generator 160 may be disposed by arranging a delay circuit for timing alignment in front of the first tap selection circuit 153. ), You can set the timing.

The coefficient seed data generating device 150 is supplied with the data of the normal equation generated for each class by the normal equation generation unit 160, and solves the normal equation for each class to obtain coefficient seed data w _{10 to} w _n9 of each class. The coefficient seed data determination unit 161 to be obtained has a coefficient seed memory 162 for storing the calculated coefficient seed data w _{10 to} w _n9 . Coefficient seed data determination section 161. In the normal equation, for example, Gauss-Jordan elimination method or the like is the pool coefficient seed data w ₁₀ ~w _n9 are obtained.

The operation of the coefficient seed data generating device 150 shown in FIG. 18 will be described. The input terminal 151 is supplied with an HD signal (525p signal or 1050i signal) as a teacher signal, and horizontal and vertical thinning processing is executed in the SD signal generation circuit 152 with respect to this HD signal, thereby providing an SD signal as an input signal. (525i signal) is generated.

In this case, when the first conversion method (obtaining a 525p signal from the 525i signal in the image signal processing unit 110 of FIG. 1) is selected, the SD signal generation circuit 152 performs thinning processing on the 525p signal to perform the SD signal. Is generated. On the other hand, when the second conversion method (the 1050i signal is obtained from the 525i signal in the image signal processing unit 110 of FIG. 1) is selected, the SD signal generation circuit 152 performs thinning processing on the 1050i signal so that the SD signal is generated. Is generated. In this case, the parameters h and v are supplied to the SD signal generation circuit 152 as a control signal, and a plurality of SD signals whose horizontal and vertical bands are changed in stages are sequentially generated.

In this SD signal 525i signal, the data (SD pixel data) of the spatial class tap selectively positioned around the target position in the HD signal (525p signal or 1050i signal) in the second tap select circuit 154 is selectively extracted. do. The second tap selection circuit 154 selects a tap based on tap position information corresponding to the selected conversion method and the motion class detected by the motion class detection circuit 158 supplied from the tap selection control circuit 156. Is executed.

The data (SD pixel data) of the spatial class tap selectively extracted by the second tap selection circuit 154 is supplied to the spatial class detection circuit 157. In this spatial class detection circuit 157, ADRC processing is performed on each SD pixel data as data of a spatial class tap to obtain a requantization code qi as class information of a spatial class (mainly class classification for waveform realization in space). (See Equation 1).

In addition, from the SD signal generated by the SD signal generation circuit 152, the data (SD pixel data) of the motion class tap located in the periphery of the position of interest in the HD signal is selectively extracted by the third tap selection circuit 155. . In this case, the third tap selection circuit 155 selects the taps based on the tap position information corresponding to the selected conversion method supplied from the tap selection control circuit 156.

The data (SD pixel data) of the motion class tap selectively extracted by the third tap selector circuit 155 is supplied to the motion class detection circuit 158. In this motion detection circuit 158, motion information MV of a motion class (mainly class classification for expressing the degree of motion) is obtained from the respective SD pixel data as the data of the motion class tap.

This motion information MV and the requantization code qi described above are supplied to the class synthesizing circuit 159. In this class synthesizing circuit 159, from these motion information MVs and the requantization code qi, a class code CL indicating a class to which the pixel data of the target position in the HD signal (525p signal or 1050i signal) belongs is obtained (mathematical formula) 3).

In addition, the data of the prediction tap (SD pixel data) located in the periphery of the position of interest in the HD signal is selectively extracted from the SD signal generated by the SD signal generation circuit 152. . In this case, the first tap select circuit 153 selects the tap based on tap position information corresponding to the selected conversion method supplied from the tap select control circuit 156.

Then, the first tap select circuit 153 selectively extracts each HD pixel data y as pixel data of the target position obtained from the HD signal supplied to the input terminal 151 and the respective HD pixel data y. From the class code CL supplied from the class synthesizing circuit 159 in correspondence with the data (SD pixel data) xi of the prediction tap and the respective HD pixel data y, the coefficient equation data w for each class in the regular equation generator 160 is obtained. A normal equation (see Equation 13) for generating _{10 to} w _n9 is generated.

Then, the coefficient seed data determination unit is that the normal equation in the pool (161) is obtained that the coefficient seed data w ₁₀ ~w _n9 for each class, the coefficient seed data w ₁₀ ~w _n9 is an address divided by each class of coefficient seed Stored in memory 162.

In this manner, the coefficient seed data generating device 150 shown in FIG. 18 generates coefficient seed data w _{10 to} w _n9 of each class stored in the information memory bank 135 of the image signal processing unit 110 of FIG. 1. Can be. In this case, the SD signal generation circuit 152 generates the SD signal 525i signal using the 525p signal or the 1050i signal by the selected conversion method. The first conversion method (the image signal processing unit 110 uses the 525i signal). Coefficient type data corresponding to a 525p signal is obtained from the signal) and the second conversion method (the 1050i signal is obtained from the 525i signal by the image signal processing unit 110).

Next, another example of the generation method of the coefficient seed data will be described. Also in this example, an example in which the coefficient seed data w _{10 to} w _n9 which are coefficient data in the above-described generation equation (5) is obtained.

20 illustrates the concept of this example. Generate a plurality of SD signals from the HD signal. For example, a total of 81 types of SD signals are generated by varying the parameters h and v for varying the horizontal and vertical bands of the filter used when generating the SD signal from the HD signal in nine steps. The learning is performed between each of the SD signals and the HD signals generated in this way to generate coefficient data Wi of the estimation equation (4). Coefficient seed data is generated using coefficient data Wi generated in correspondence with each SD signal.

First, a method of obtaining coefficient data of the estimation equation will be described. Here, an example will be described in which coefficient data Wi (i = 1 to n) of the estimation equation of Expression 4 is obtained by the least square method. As a generalized example, the observation equation of Equation (15) is considered with X as input data, W as coefficient data, and Y as prediction values. In this equation (15), m represents the number of training data and n represents the number of prediction taps.                 

The least-squares method is applied to the data collected by the observation equation (15). Based on this observation equation (15), the residual equation (16) can be obtained.

An optimum value of each Wi from the residual equation of Equation 16, it is conceivable a case in which a condition that minimizes the e ² in Equation (17) holds. That is, the condition of Equation 18 may be considered.

In other words, W ₁ , W ₂ ,... Calculate W _n . Therefore, equation (19) can be obtained from the residual equation of equation (16). Then, equations (20) can be obtained from equations (19) and (15).

Then, from the equations (16) and (20), the normal equation of equation (21) can be obtained.                 

Since the regular equation of Equation 21 can formulate the same number of equations as the unknown number n, the optimum value of each Wi can be obtained. In this case, the system of equations is solved using Gaussian-Jordan elimination.

Next, a method of obtaining coefficient seed data using coefficient data generated corresponding to each SD signal will be described.

_Assume that the coefficient data of any class by learning using the SD signals corresponding to the parameters h and v is k _vhi . Where i is the number of prediction taps. _Obtain the coefficient seed data of this class from this k _vhi .

Each coefficient data Wi (i = 1 to n) is expressed by the above expression (5) using the coefficient seed data w _{10 to} w _n9 . Here, considering the use of the least squares method for the coefficient data Wi, the residual is represented by equation (22).

Here, t _j is shown in Equation 7 above. By applying the least squares method to (22), (23) can be obtained.

Here, when X _jk and Y _j are defined as in Equation 24 and Equation 25, Equation 23 can be rewritten as in Equation 26. This equation (26) is also a normal equation, and the coefficient seed data w _{10 to} w _n9 can be calculated by solving this equation with a general solution such as a Gaussian-Jordan cancellation method.

FIG. 21 shows a configuration of the coefficient seed data generating device 150 'that generates coefficient seed data based on the concept shown in FIG. In Fig. 21, parts corresponding to those in Fig. 20 are denoted by the same reference numerals and detailed description thereof will be omitted.

The coefficient seed data generating device 150 'selects each of the HD pixel data y as the pixel data of the target position obtained from the HD signal supplied to the input terminal 151 and the first tap corresponding to the HD pixel data y, respectively. The coefficient data Wi (for each class) is derived from the data (SD pixel data) xi of the prediction tap selectively extracted by the circuit 153 and the class code CL output from the class synthesizing circuit 159 corresponding to each HD pixel data y, respectively. A normal equation generating unit 171 generates a normal equation (see Equation 21) for obtaining i = 1 to n).

In this case, the training data is generated from a combination of one HD pixel data y and n predicted tap pixel data corresponding thereto, and the parameters h and v to the SD signal generation circuit 152 are sequentially changed so that horizontal and vertical A plurality of SD signals whose bands have been changed step by step are sequentially generated, and generation of learning data is performed between the HD signal and each SD signal, respectively. Accordingly, the normal equation generating unit 171 generates a normal equation for obtaining coefficient data Wi (i = 1 to n) for each class corresponding to each of the SD signals.

The coefficient seed data generating device 150 'is supplied with the data of the normal equation generated by the normal equation generation unit 171, and solves the normal equation to obtain coefficient data Wi of each class corresponding to each SD signal. The coefficient data determination unit 172 and a normal to generate a normal equation (see Equation 26) for obtaining coefficient seed data w _{10 to} w _n9 for each class using the coefficient data Wi of each class corresponding to each SD signal. It has an equation generator 173.

The coefficient seed data generating device 150 'is supplied with the data of the normal equations generated for each class by the normal equation generation unit 173, and solves the normal equations for each class to calculate the coefficient seed data w _{10 to} w _n9 of each class. And a coefficient seed data determining unit 174 for obtaining? And a coefficient seed memory 162 for storing the obtained coefficient seed data w _{10 to} w _n9 .

Other than the coefficient seed data generating device 150 'shown in FIG. 21, it is comprised similarly to the coefficient seed data generating device 150 shown in FIG.

The operation of the coefficient seed data generating device 150 'shown in FIG. 21 will be described. The input terminal 151 is supplied with an HD signal (525p signal or 1050i signal) as a teacher signal, and horizontal and vertical thinning processing is executed in the SD signal generation circuit 152 with respect to this HD signal, thereby providing an SD signal as an input signal. (525i signal) is generated.

In this case, when the first conversion method (obtaining a 525p signal from the 525i signal in the image signal processing unit 110 of FIG. 1) is selected, the SD signal generation circuit 152 performs thinning processing on the 525p signal to perform the SD signal. Is generated. On the other hand, when the second conversion method (obtaining a 1050i signal from the 525i signal in the image signal processing unit 110 of FIG. 1) is selected, the SD signal generation circuit 152 performs thinning processing on the 1050i signal to generate an SD signal. Is generated. In this case, the parameters h and v are supplied to the SD signal generation circuit 152 as a control signal, and a plurality of SD signals whose horizontal and vertical bands are changed in stages are sequentially generated.

From this SD signal 525i signal, the data (SD pixel data) of the spatial class tap selectively located in the periphery of the target position in the HD signal (525p signal or 1050i signal) in the second tap selection circuit 154 is selectively extracted. do. The second tap selection circuit 154 selects a tap based on tap position information corresponding to the selected conversion method and the motion class detected by the motion class detection circuit 158 supplied from the tap selection control circuit 156. Is executed.

The data (SD pixel data) of the spatial class tap selectively extracted by the second tap selection circuit 154 is supplied to the spatial class detection circuit 157. In this space class detection circuit 157, ADRC processing is performed on each SD pixel data as data of a space class tap to obtain a requantization code qi as class information of a space class (mainly class classification for waveform representation in space). (See Equation 1).

Further, data of the motion class taps (SD pixel data) located in the periphery of the point of interest in the HD signal by the third tap selection circuit 155 is selectively extracted from the SD signal generated by the SD signal reproduction circuit 152. do. In this case, the tap selection is performed in the third tap selection circuit 155 based on the tap position information corresponding to the selected conversion method supplied from the tap selection control circuit 156.

The data (SD pixel data) of the motion class tap selectively extracted by the third tap selector circuit 155 is supplied to the motion class detection circuit 158. In this motion class detection circuit 158, motion information MV of a motion class (mainly class classification for expressing the degree of motion) is obtained from the respective SD pixel data as the data of the motion class tap.

This motion information MV and the requantization code qi described above are supplied to the class synthesizing circuit 159. In this class synthesizing circuit 159, from these motion information MVs and the requantization code qi, a class code CL indicating a class to which the pixel data of the target position in the HD signal (525p signal or 1050i signal) belongs is obtained (mathematical formula) 3).

In addition, the data of the prediction tap (SD pixel data) located in the periphery of the position of interest in the HD signal is selectively extracted from the SD signal generated by the SD signal generation circuit 152. . In this case, the first tap select circuit 153 selects the tap based on tap position information corresponding to the selected conversion method supplied from the tap select control circuit 156.

Then, the first tap select circuit 153 selectively extracts each HD pixel data y as pixel data of the target position obtained from the HD signal supplied to the input terminal 151 and the respective HD pixel data y. From the class code CL outputted from the class synthesizing circuit 159 corresponding to each HD pixel data y xi of the prediction tap data xi and the HD pixel data y, the normal equation generating unit 171 uses the SD signal generating circuit 152. A regular equation (see Equation 21) for obtaining coefficient data Wi (i = 1 to n) is generated for each class corresponding to each of the SD signals generated in Eq.

The normal equation is solved by the coefficient data determination unit 172 to obtain coefficient data Wi of each class corresponding to each SD signal. The normal equation generating unit 173 generates a normal equation (see Equation 26) for obtaining the coefficient seed data w _{10 to} w _n9 for each class from the coefficient data Wi of each class corresponding to each SD signal.

Then, the coefficient seed data, the normal equation from the determiner 174 is the pool is obtained that the coefficient seed data w ₁₀ ~w _n9 for each class, the coefficient seed data w ₁₀ ~w _n9 is the coefficient seed memory address divided by class Stored at 162.

As described above, in the coefficient seed data generating device 150 ′ shown in FIG. 21, coefficient seed data w _{10 to} w _n9 of each class stored in the information memory bank 135 of the image signal processing unit 110 of FIG. 1 are generated. can do. In this case, the SD signal generation circuit 152 generates the SD signal 525i signal using the 525p signal or the 1050i signal by the selected conversion method. The first conversion method (the image signal processing unit 110 uses the 525i signal from the 525i signal). Coefficient type data corresponding to a 525p signal) and a second conversion method (the image signal processing unit 110 obtains a 1050i signal from the 525i signal) may be generated.

In the image signal processing unit 110 of FIG. 1, the generation equation (5) is used to generate the coefficient data Wi (i = 1 to n). For example, the equation (27), equation (28), or the like may be used. Not only can it be realized, but it can also be realized by expressions of different polynomials or other functions.

In addition, in the image signal processing unit 110 of FIG. 1, the parameter h specifying the horizontal resolution and the parameter v specifying the vertical resolution are set, and the horizontal and vertical resolutions of the image are adjusted by adjusting these parameters h and v values. Although it can be shown, the thing which can adjust the noise removal degree of an image similarly can be comprised by setting the parameter z which specifies the noise removal degree (noise reduction degree), for example, and adjusting this parameter z value.                 

In this case, as a generation equation for generating coefficient data Wi (i = 1 to n), for example, equation 29, equation 30, and the like can be used, and the expressions are expressed by polynomials or other functions having different orders. Can also be realized.

As described above, the coefficient seed data that is the coefficient data of the generation equation including the parameter z is similar to the case of generating the coefficient seed data that is the coefficient data of the generation equation including the parameters h and v described above. The coefficient seed data generating device 150 shown in FIG. 21 or the coefficient seed data generating device 150 'shown in FIG. 21 can be generated.

In this case, the parameter z is supplied to the SD signal generation circuit 152 as a control signal, and the noise addition state to the SD signal is varied step by step when generating the SD signal from the HD signal corresponding to the parameter z value. Thus, by registering the training data by varying the noise addition state to the SD signal in stages, the coefficient seed data for obtaining continuous noise removal can be generated.

Here, some examples of the details of the noise adding method corresponding to the parameter z value are shown.

For example, as shown in Fig. 22A, an SD signal whose noise level changes in steps is added to the SD signal by adding a noise signal having the amplitude level changed in steps.

For example, as shown in FIG. 22B, a noise signal having a constant amplitude level is added to the SD signal, and the screen area to be added is varied in stages.

For example, as shown in Fig. 22C, an SD signal (for one screen) is prepared which does not contain noise and which contains noise. When generating a normal equation, a plurality of learnings are performed on each SD signal.

For example, in "Noise 0", 100 lessons are performed on an SD signal that is no noise, and "Noise i" is performed by 30 lessons for an SD signal that is noisy and in addition to the noisy SD signal. Perform 70 lessons. In this case, "noise i" becomes a learning system for calculating coefficient seed data with high degree of noise removal. As described above, coefficient type data for obtaining continuous noise removal can be obtained by changing the number of times of learning for the noise-free and noisy SD signal in a stepwise manner.

Although this technique is not described in detail, it can be realized in the form of addition of a regular equation.

First, learning to calculate coefficient data of the estimation equation in "noise 0" to "noise i" is performed. The normal equation at this time is as shown in Equation 21 above. Here, if P _ij and Q _j are defined as in Equation 31 and 32, respectively, Equation 21 can be rewritten as in Equation 33. Here, x _ij is the i-th learning value of the SD pixel data of the j-th prediction tap position, y _i is the i-th learning value of the HD pixel data, and W _i is a coefficient.

In this case, the left side of Equation 33 is defined as [P1 _ij ] and the right side as [Q1 _i ]. The left side of Equation 33 is defined as [P2 _ij ], and the right side is defined as [Q2 _i ]. In this case, [P _aij ] and [Q _ai ] are defined as in Equations 34 and 35. However, a is 0≤a≤1.

Here, the normal equation in the case of a = 0 is expressed by the equation (36), which is equivalent to the normal equation in the case of "noise 0" in Fig. 22C, and in the case of "noise i" in the case of a = 0.7, It is equivalent to the equation.

By changing this a stepwise to create a normal equation for each noise level, the target coefficient seed data can be obtained. In this case, as described in the coefficient seed data generating device 150 'of FIG. 21, coefficient data Wi is calculated from the normal equation of each noise level, and coefficient seed data can be obtained using the coefficient data of each step. have.

In addition, by combining the normal equation of each noise level, it is also possible to generate a normal equation for obtaining coefficient seed data as shown in Equation (13). This technique is described in detail below. Here, a case of generating a normal equation for obtaining coefficient seed data using Equation (30) is considered.

SD signals of noise levels corresponding to some kinds of parameters z are generated in advance, and learning is performed, and [P] and [Q] represented by Equations 34 and 35 are prepared. These are denoted as [P _nij ] [Q _ni ]. In addition, Equation 7 may be rewritten as in Equation 37.

In this case, Equations 24 and 25 are rewritten as Equations 38 and 39, respectively. The coefficient seed data w _ij can be obtained by solving Equation 40 with respect to these equations. Here, the variable representing the total number of prediction taps can be rewritten as m.

In addition, in the image signal processing unit 110 of FIG. 1, the parameter h specifying the horizontal resolution and the parameter v specifying the vertical resolution are set, and the horizontal and vertical resolutions of the image are adjusted by adjusting these parameters h and v values. Although it can be shown, it can also be configured to adjust the horizontal and vertical resolution to one parameter. For example, one parameter r that specifies horizontal and vertical resolutions is set. In this case, for example, r = 1 is h = 1, v = 1, r = 2 is h = 2, v = 2,... Or r = 1 for h = 1, v = 2, r = 2 for h = 2, v = 3,... The correspondence relationship is as follows. In this case, a polynomial of r or the like is used as a generation expression for generating coefficient data Wi (i = 1 to n).

In addition, in the image signal processing unit 110 of FIG. 1, the horizontal and vertical resolution of an image is set by setting a parameter h specifying a horizontal resolution and a parameter v specifying a vertical resolution, and adjusting these multiple types of parameters h and v values. Although it can be shown that the parameter r which designates the horizontal and vertical resolution mentioned above and the parameter z which designates the said noise removal degree (noise reduction degree) are set, these multiple types of parameter r and z value are set. It is also possible to adjust the horizontal and vertical resolution of the image and the degree of noise removal by adjusting.

In this case, as a generation expression for generating coefficient data Wi (i = 1 to n), for example, not only Equation 41 can be used but also the expression expressed by a polynomial or another function having different orders. .

In this manner, a user interface for adjusting a plurality of types of parameters r and z can also be configured as shown in FIG. The user can move the position of the icon 115a on the adjustment screen 115 by manipulating the joystick 200a, so that the parameter r value specifying the resolution and the parameter z value specifying the noise removal degree (noise reduction) are adjusted. It can be adjusted arbitrarily. 24 shows an enlarged view of a portion of the adjustment screen 115. The parameter r value for determining the resolution by adjusting the icon 115a to the left and right is adjusted, and the parameter z value for determining the degree of noise removal by adjusting the icon 115a to move up and down.

The user can adjust the parameters r and z with reference to the adjustment screen 115 displayed on the display 111 (refer FIG. 2), and the adjustment can be made easy. The parameter r and z values adjusted by the user may be numerically displayed on the adjustment screen 115.

Thus, the coefficient type data which is the coefficient data of the generation formula including the parameters r and z is the same as the case of generating the coefficient type data which is the coefficient data of the generation formula including the above-described parameters h and v. It can be generated by the coefficient seed data generating device 150 or the coefficient seed data generating device 150 'shown in FIG.

In this case, the parameters r and z are supplied to the SD signal generating circuit 152 as a control signal, and the horizontal and vertical bands of the SD signal when generating the SD signal from the HD signal corresponding to the values of the parameters r and z, The noise adding state to the SD signal is varied in stages.

23 shows an example of generation of SD signals corresponding to the parameters r and z values. In this example, the parameters r and z are varied in nine steps, respectively, and a total of 81 types of SD signals are generated. The parameters r and z may be varied in more than nine steps. In this case, the accuracy of the calculated coefficient seed data is improved, but the calculation amount is increased.

In addition, in the image signal processing unit 110 of FIG. 1, the horizontal and vertical resolution of an image is set by setting a parameter h specifying a horizontal resolution and a parameter v specifying a vertical resolution, and adjusting these multiple types of parameters h and v values. In addition to these parameters h and v, the parameter z specifying the noise reduction degree (noise reduction degree) described above is set, and the plurality of types of parameters h and v are adjusted to adjust the horizontal Similarly, the vertical resolution and noise rejection can be adjusted.

In this case, as a generation formula for generating coefficient data Wi (i = 1 to n), for example, not only Equation (42) can be used but also a formula expressed by a polynomial or a different function having different orders. .                 

Thus, the user interface for adjusting the parameters h, v, z is also configured as shown in FIG. The user can move the position of the icon 115a on the adjustment screen 115 by manipulating the joystick 200a, and the parameters h and v values for specifying the resolution and the parameters for specifying noise reduction (noise reduction). The z value can be adjusted arbitrarily.

26 shows an enlarged portion of the adjustment screen 115. By moving the icon 115a from side to side, the parameter h value for determining the horizontal resolution is adjusted, while the parameter v value for determining the vertical resolution by the icon 115a is moved up and down, and the icon 115a is adjusted. By moving inwardly, the parameter z value for determining the noise removal degree is adjusted. In order to move the icon 115a inward, for example, the joystick 200a may be operated in an oblique direction.

In this case, the inward direction can be realized by changing the size of the icon 115a, the density of the color, the color tone, or the like. The icon 115a shown by the broken line in the figure shows the state in which the icon 115a shown by the solid line is changed in the inside direction by changing the size.

The user can adjust the parameters h, v, z with reference to the adjustment screen 115 displayed on the display 111 (refer FIG. 2), and the adjustment can be made easy. The parameter h, v, z values adjusted by the user may be displayed numerically on the adjustment screen 115.

Thus, the coefficient seed data which is the coefficient data of the generation formula including the parameters h, v and z is similar to the case of generating the coefficient seed data which is the coefficient data of the generation expression including the above-described parameters h and v. The coefficient seed data generating device 150 shown in FIG. 21 or the coefficient seed data generating device 150 'shown in FIG. 21 can be generated.                 

In this case, the parameters h, v, z are supplied to the SD signal generation circuit 152 as a control signal, and when the SD signal is generated from the HD signal corresponding to the values of the parameters h, v, z, The vertical band and the noise adding state to the SD signal vary in stages.

25 shows an example of generating an SD signal corresponding to the parameter h, v, z values. In this example, the parameters h, v, and z are each varied in nine steps, and a total of 729 types of SD signals are generated. The parameters h, v, and z may be varied in more than nine steps. In this case, the accuracy of the calculated coefficient seed data is improved, but the calculation amount is increased.

The processing in the image signal processing unit 110 in FIG. 1 can also be implemented in software by the image signal processing apparatus 300 as shown in FIG. 27, for example.

First, the image signal processing apparatus 300 shown in FIG. 27 will be described. The image signal processing apparatus 300 includes a CPU 301 for controlling the operation of the entire apparatus, a ROM (Read Only Memory) 302 in which operation programs, coefficient seed data, etc. of the CPU 301 are stored, and a CPU ( It has a random access memory (RAM) 303 constituting the working environment of 301. These CPUs 301, ROM 302, and RAM 303 are connected to the bus 304, respectively.

The image signal processing apparatus 300 further includes a hard disk drive (HDD) 305 as an external storage device and a floppy (R) disk drive (FDD) 307 for driving the floppy (R) disk 306. have. These drives 305 and 307 are connected to the bus 304, respectively.                 

In addition, the image signal processing apparatus 300 includes a communication unit 308 that connects to a communication network 400 such as the Internet by wire or wirelessly. This communication unit 308 is connected to the bus 304 via the interface 309.

In addition, the image signal processing apparatus 300 includes a user interface unit. This user interface unit has a remote control signal receiving circuit 310 for receiving a remote control signal RM from the remote control transmitter 200 and a display 311 made of a liquid crystal display (LCD) or the like. The receiving circuit 310 is connected to the bus 304 via the interface 312, and likewise the display 311 is connected to the bus 304 via the interface 313.

The image signal processing apparatus 300 also has an input terminal 314 for inputting an SD signal and an output terminal 315 for outputting an HD signal. Input terminal 314 is connected to bus 304 via interface 316, and output terminal 315 is likewise connected to bus 304 via interface 317.

As described above, instead of storing the processing program, the coefficient seed data and the like in the ROM 302 in advance, for example, it is downloaded from the communication network 400 such as the Internet via the communication unit 308, and the hard disk or RAM 303 Can be accumulated and used. Further, these processing programs, coefficient seed data, and the like may be provided by the floppy (R) disk 306.

Instead of inputting the SD signal to be processed from the input terminal 314, it may be recorded in advance on the hard disk or may be downloaded from the communication network 400 such as the Internet via the communication unit 308. In addition, instead of outputting the processed HD signal to the output terminal 315, in parallel with it, it is supplied to the display 311 to display an image, to store it on a hard disk, or to a communication network such as the Internet through the communication unit 308. You may make it send to (400).

A processing procedure for obtaining an HD signal from an SD signal in the image signal processing apparatus 300 shown in FIG. 27 will be described with reference to the flowchart in FIG. 28.

First, processing is started in step ST1, and SD pixel data is input in frame units or field units in step ST2. When the SD pixel data is input from the input terminal 314, the SD pixel data is temporarily stored in the RAM 303. When the SD pixel data is recorded on the hard disk, the SD pixel data is read by the hard disk drive 305 and temporarily stored in the RAM 303. Then, in step ST3, it is determined whether or not the processing of all frames or all fields of the input SD pixel data is finished. When the process ends, the process ends in step ST4. On the other hand, when the process has not ended, the process proceeds to step ST5.

In this step ST5, the image quality specification values (for example, the parameters h, v values, etc.) input by the user by operating the remote control transmitter 200 are read from the RAM 303, for example. Then, in step ST6, the coefficient data Wi of the estimated equation (see Equation 4) of each class is generated by the generation equation (for example, Equation 5) using the read-quality designated value and the coefficient seed data of each class. Create

Subsequently, in step ST7, from the SD pixel data input in step ST2, pixel data of a class tap and a prediction tap are obtained corresponding to each HD pixel data to be generated. In step ST8, it is determined whether or not the processing for obtaining the HD pixel data in the entire area of the inputted SD pixel data has ended. When it is finished, the process returns to step ST2 and the process shifts to input processing of the SD pixel data of the next frame or field. On the other hand, when the process has not ended, the process proceeds to step ST9.

In this step ST9, the class code CL is generated from the SD pixel data of the class tap acquired in step ST7. Then, in step ST10, HD pixel data is generated by the estimation equation using the coefficient data corresponding to the class code CL and the SD pixel data of the prediction tap. Then, the process returns to step ST7 and repeats the same processing as described above. .

By performing the processing according to the flowchart described in FIG. 28 as described above, the HD pixel data constituting the HD signal can be obtained by processing the SD pixel data constituting the input SD signal. As described above, the HD signal obtained by this processing is output to the output terminal 315 or supplied to the display 311 to display an image therefrom, as well as to the hard disk drive 305 to be supplied to the hard disk. Is recorded.

In addition, although illustration of the processing apparatus is abbreviate | omitted, the process by the coefficient seed data generating apparatus 150 of FIG. 18 can also be implemented by software.

A processing procedure for generating coefficient seed data will be described with reference to the flowchart in FIG. 29.

First, processing is started in step ST21, and an image quality pattern (e.g., specified by parameters h and v) used for learning is selected in step ST22. In step ST23, it is determined whether or not learning has been completed for the entire picture quality pattern. When the learning is not finished for the entire picture quality selection pattern, the flow advances to step ST24.

In this step ST24, the known HD pixel data is input in units of frames or fields. Then, in step ST25, it is determined whether or not the processing is finished for all the HD pixel data. When it is finished, the process returns to step ST22 to select the next image quality pattern and repeats the same process as described above. On the other hand, when it is not complete | finished, it progresses to step ST26.

In this step ST26, SD pixel data is generated based on the image quality pattern selected in step ST22 from the HD pixel data input in step ST24. In step ST27, the pixel data of the class tap and the prediction tap are obtained from the SD pixel data generated in step ST26 corresponding to the HD pixel data input in step ST24. In step ST28, it is determined whether or not the learning process has ended in the entire area of the generated SD pixel data. When the learning process is finished, the process returns to step ST24 to execute input of the next HD pixel data to repeat the same process as described above, and when the learning process is not finished, the process goes to step ST29.

In this step ST29, the class code CL is generated from the SD pixel data of the class tap acquired in step ST27. In step ST30, a normal equation (see Equation 13) is generated. Thereafter, the process returns to step ST27.

In addition, in step ST27, when learning is complete about all image quality data, it progresses to step ST31. In this step ST31, coefficient type data of each class is calculated by solving a normal equation by a Gaussian-Jordan erasing method or the like. In step ST32, the coefficient type data is stored in a memory, and then the processing is terminated in step ST33.

As described above, by processing according to the flowchart illustrated in FIG. 29, the coefficient seed data of each class can be obtained by the same technique as that of the coefficient seed data generating device 150 shown in FIG. 18.

Although the illustration of the processing device is omitted, the processing in the coefficient seed data generating device 150 'in FIG. 21 can also be realized by software.

A processing procedure for generating coefficient seed data will be described with reference to the flowchart in FIG. 30.

First, processing is started in step ST41, and an image quality pattern (e.g., specified by parameters h and v) used for learning is selected in step ST42. Then, in step ST43, it is determined whether or not the calculation processing of the coefficient data for the entire picture quality pattern is finished. If it is not finished, the flow advances to step ST44.

In this step ST44, the known HD pixel data is input in units of frames or fields. Then, in step ST45, it is determined whether or not the processing is finished for all the HD pixel data. When it is not finished, in step ST46, SD pixel data is generated from the HD pixel data input in step ST44 based on the picture quality pattern selected in step ST42.

In step ST47, the pixel data of the class tap and the prediction tap are obtained from the SD pixel data generated in step ST46 corresponding to the HD pixel data input in step ST44. In step ST48, it is determined whether or not the learning process has ended in the entire area of the generated SD pixel data. When the learning process is finished, the process returns to step ST44 to execute input of the next HD pixel data to repeat the same process as described above, and to proceed to step ST49 when the learning process is not finished.

In step ST49, the class code CL is generated from the SD pixel data of the class tap acquired in step ST47. In step ST50, a normal equation (see Equation 21) for obtaining coefficient data is generated. Thereafter, the process returns to step ST47.

When the processing is completed for all the HD pixel data in step ST45 described above, in step ST51, the normal equation generated in step ST50 is solved by a Gaussian-Jordan erasure method or the like to calculate coefficient data of each class. After that, the process returns to step ST42, the next image quality pattern is selected, the same processing as described above is repeated, and coefficient data of each class corresponding to the next image quality pattern is obtained.

In addition, in the above-mentioned step ST43, when the calculation process of the coefficient data with respect to the whole picture quality pattern is complete | finished, it progresses to step ST52. In this step ST52, a regular equation (see Equation 26) for calculating coefficient seed data is generated from the coefficient data for the entire picture quality pattern.

In step ST53, coefficient type data of each class is calculated by solving the regular equations generated in step ST52 by the Gaussian-Jordan erasure method, etc., and storing the coefficient seed data in memory in step ST54, and then in step ST55 The process ends.

By processing in accordance with the flowchart described in FIG. 30 in this manner, the coefficient seed data of each class can be obtained by the same technique as that of the coefficient seed data generating device 150 'shown in FIG.

Next, another embodiment of the present invention will be described. 31 shows the configuration of a TV receiver 100 'as another embodiment. The TV receiver 100 'also obtains a 525i signal as an SD signal from a broadcast signal, converts the 525i signal into a 525p signal or a 1050i signal as an HD signal, and displays an image by the 525p signal or 1050i signal. In Fig. 31, parts corresponding to those in Fig. 1 are denoted by the same reference numerals.

In the TV receiver 100 ', the image signal processing unit 110 of the TV receiver 100 shown in FIG. 1 is replaced with the image signal processing unit 110', and performs the same operation as that of the TV receiver 100. FIG.

The image signal processing unit 110 'will be described in detail. In the image signal processing unit 110 ', portions corresponding to those of the image signal processing unit 110 shown in Fig. 1 are denoted by the same reference numerals, and detailed description thereof will be omitted.

This image signal processing section 110 'has an information memory bank 135'. In the information memory bank 135 ', similar to the information memory bank 135 in the image signal processing unit 110 shown in FIG. 1, operation designation information for storing in the register 130, and the registers 131 to 131. The tap position information for storing in 133 is stored in advance. In this information memory bank 135 ', coefficient data for each combination of the class and the parameters h and v values corresponding to each of the first conversion method 525p and the second conversion method 1050i is stored in advance. have. The generation method of this coefficient data is mentioned later.

The operation of this image signal processing unit 110 'will be described.

From the SD circuit 525i signal stored in the buffer memory 109, the data (SD pixel data) of the spatial class taps are selectively extracted by the second tap selection circuit 122. In this case, the second tap selection circuit 122 selects the tap based on the conversion method selected by the user supplied from the register 132 and the tap position information corresponding to the motion class detected by the motion class detection circuit 125. This is done.

The data (SD pixel data) of the spatial class tap selectively extracted by the second tap selection circuit 122 is supplied to the spatial class detection circuit 124. In this spatial class detection circuit 124, ADRC processing is performed on each SD pixel data as data of a spatial class tap to obtain a requantization code qi as class information of a spatial class (mainly class classification for waveform representation in space). (See Equation 1).

In addition, data of the motion class tap (SD pixel data) is selectively extracted from the SD signal 525i signal stored in the buffer memory 109 by the third tap selection circuit 123. In this case, the third tap selection circuit 123 selects the taps based on the tap position information corresponding to the conversion method selected by the user, supplied from the register 133.                 

The data (SD pixel data) of the motion class tap selectively extracted by the third tap selection circuit 123 is supplied to the motion class detection circuit 125. In this motion class detection circuit 125, motion information MV of a motion class (mainly class classification for indicating the degree of motion) is obtained from the respective SD pixel data as the data of the motion class tap.

This motion information MV and the requantization code qi described above are supplied to the class synthesizing circuit 126. In this class synthesizing circuit 126, from these motion information MVs and the requantization code qi, a class code CL indicating a class to which the pixel data (pixel data at the target position) of the HD signal (525p signal or 1050i signal) to be created belongs. Is obtained (see Equation 3). This class code CL is supplied to the coefficient memory 134 as read address information.

In the coefficient memory 134, coefficient data of each class corresponding to the parameter h and v values adjusted by the user and the conversion method are loaded and stored from the information memory bank 135 ', for example, in the vertical blanking period. . As described above, the class code CL is supplied as the read address information, whereby coefficient data Wi corresponding to the class code CL is read out from this coefficient memory 134 and supplied to the estimation prediction calculating circuit 127.

When coefficient data corresponding to the adjusted parameters h and v values are not accumulated in the information memory bank 135 ', the coefficient data corresponding to the values before and after the adjusted parameters h and v values are displayed. It is also possible to obtain coefficient data corresponding to the adjusted parameters h and v values by reading from the memory bank 135 'and using the interpolation arithmetic processing using them.

Further, data of the prediction tap (SD pixel data) is selectively extracted from the SD signal 525i signal stored in the buffer memory 109 by the first tap selection circuit 121. In this case, the first tap selection circuit 121 selects the tap based on the tap position information corresponding to the conversion method selected by the user, supplied from the register 131. The data (SD pixel data) xi of the prediction tap selectively extracted by the first tap selection circuit 121 is supplied to the estimation prediction calculating circuit 127.

In the estimation prediction calculating circuit 127, from the data (SD pixel data) xi of the prediction tap and the coefficient data Wi read out from the coefficient memory 134, the pixel data of the HD signal to be created, that is, the pixel data of the target position ( HD pixel data) y is calculated (see Equation 4). In this case, data of four pixels constituting the HD signal are simultaneously generated.

Accordingly, when the first conversion method for outputting the 525p signal is selected, in the odd (o) field and the even (e) field, the line data L1 at the same position as the line of the 525i signal and the upper and lower lines of the 525i signal are Line data L2 at the intermediate position is generated (see FIG. 4). When the second conversion method for outputting the 1050i signal is selected, in the odd (o) field and the even (e) field, the line data L1, L1 'and the line of the 525i signal are located close to the line of the 525i signal. Line data L2, L2 'at a position far from is generated (see Fig. 5).

The line data L1 and L2 (L1 ', L2') generated in the estimation prediction calculating circuit 127 are supplied to the line sequential converting circuit 129. In this line sequential conversion circuit 129, the line data L1 and L2 (L1 ', L2') are line-sequentially generated to generate an HD signal. In this case, the line sequential conversion circuit 129 operates according to the operation instruction information corresponding to the conversion method selected by the user, supplied from the register 130. Therefore, when the first conversion method 525p is selected by the user, the 525p signal is output from the line sequential conversion circuit 129. On the other hand, when the second conversion method 1050i is selected by the user, the 1050i signal is output from the line sequential conversion circuit 129.

In this manner, in the image signal processing unit 110 ', coefficient data Wi (i = 1 to n) of the estimated equation corresponding to the adjusted parameters h and v values is used, and HD pixel data y is calculated. Therefore, the user can freely adjust the image quality of the image by the HD signal on the axis of horizontal resolution and vertical resolution by adjusting the parameters h and v values.

As described above, coefficient data for each combination of class and parameter h and v values corresponding to each of the first and second conversion methods is stored in the information memory bank 135 'in advance. This coefficient data is generated by learning in advance.

In the above description, as another example of the generation method of the coefficient seed data, first, for each SD signal obtained by varying the parameter h and v values stepwise, coefficient data of each class is generated by learning using the same, and then for each SD signal. Using coefficient data of each class in the above, obtaining coefficient coefficient data of each class was described. The coefficient data for each combination of the class and the parameters h and v values, which are stored in the information memory bank 135 'in advance, can be generated in the same manner as in the first half of the method for generating the coefficient seed data.                 

32 shows a coefficient data generating device 180. In the coefficient data generating device 180, parts corresponding to those of the coefficient type data generating device 150 'shown in Fig. 21 are denoted by the same reference numerals, and detailed description thereof is omitted.

The coefficient data generating device 180 has a coefficient seed memory 162. In this coefficient type memory 162, coefficient data Wi of each class corresponding to each SD signal determined by the coefficient data determination unit 172 is stored. The rest of the coefficient data generating device 180 is configured similarly to the coefficient seed data generating device 150 'shown in FIG.

The operation of the coefficient data generating device 180 shown in FIG. 32 will be described. The input terminal 151 is supplied with an HD signal (525p signal or 1050i signal) as a teacher signal, and horizontal and vertical thinning processing is performed by the SD signal generation circuit 152 with respect to this HD signal, thereby providing an SD signal as an input signal. (525i signal) is generated.

In this case, when the first conversion method (obtaining a 525p signal from the 525i signal in the image signal processing unit 110 'in Fig. 31) is selected, the SD signal generation circuit 152 performs thinning processing on the 525p signal and performs SD. The signal is generated. On the other hand, when the second conversion method (obtaining a 1050i signal from the 525i signal in the image signal processing unit 110 'of FIG. 31) is selected, the SD signal generation circuit 152 performs thinning processing on the 1050i signal to perform the SD signal. Is generated. In this case, the parameters h and v are supplied to the SD signal generation circuit 152 as a control signal, and a plurality of SD signals in which the horizontal and vertical bands are gradually changed are sequentially generated.                 

From this SD signal 525i signal, in the second tap selector circuit 154, the data (SD pixel data) of the spatial class tap located around the target position in the HD signal (525p signal or 1050i signal) is selectively Extracted. In the second tap selection circuit 154, the tap is selected based on the selected conversion method supplied from the tap selection control circuit 156 and the tap position information corresponding to the motion class detected by the motion class detection circuit 158. Is selected.

The data (SD pixel data) of the spatial class tap selectively extracted by the second tap selection circuit 154 is supplied to the spatial class detection circuit 157. In this space class detection circuit 157, ADRC processing is performed on each SD pixel data as data of a space class tap to obtain a requantization code qi as class information of a space class (see Equation 1).

In addition, from the SD signal generated by the SD signal generation circuit 152, the data (SD pixel data) of the motion class tap located in the periphery of the target position in the HD signal in the third tap selection circuit 155 is selectively Extracted. In this case, in the third tap selection circuit 155, tap selection is performed based on tap position information corresponding to the selected conversion method supplied from the tap selection control circuit 156.

The data (SD pixel data) of the motion class tap selectively extracted by the third tap selector circuit 155 is supplied to the motion class detection circuit 158. In this motion class detection circuit 158, motion information MV of a motion class is obtained from each SD pixel data as data of a motion class tap.

This motion information MV and the requantization code qi described above are supplied to the class synthesizing circuit 159. In this class synthesizing circuit 159, from these motion information MVs and the requantization code qi, a class code CL indicating a class to which the pixel data of the target position in the HD signal (525p signal or 1050i signal) belongs is obtained (mathematical formula) 3).

In addition, from the SD signal generated by the SD signal generation circuit 152, the data (SD pixel data) of the prediction tap located in the periphery of the point of interest in the HD signal is selectively extracted by the first tap selection circuit 153. do. In this case, in the first tap selection circuit 153, tap selection is performed based on tap position information corresponding to the selected conversion method supplied from the tap selection control circuit 156.

Then, the first tap select circuit 153 selectively extracts each HD pixel data y as pixel data of the target position obtained from the HD signal supplied to the input terminal 151 and the respective HD pixel data y. From the class code CL output from the class synthesizing circuit 159 in correspondence with the data (SD pixel data) xi of the prediction tap and the respective HD pixel data y, the normal equation generating unit 171 uses the SD signal generating circuit 152. Corresponding to each of the SD signals generated in Fig. 2), a regular equation (see Equation 21) for obtaining the coefficient data Wi (i-1 to n) is generated for each class.

The regular equation is solved by the coefficient data determination unit 172, and coefficient data Wi of each class corresponding to each SD signal is obtained. That is, the coefficient data determination unit 172 obtains coefficient data Wi for each combination of the class and the parameters h and v values. The coefficient data Wi is stored in the coefficient type memory 162 divided into addresses for each combination of class and parameter h and v values.

As described above, in the coefficient data generation device 180 shown in FIG. 32, each combination of the class and the parameter h and v values stored in the information memory bank 135 'of the image signal processing unit 110' of FIG. Coefficient data Wi can be generated. In this case, the SD signal generation circuit 152 generates an SD signal 525i signal using a 525p signal or a 1050i signal by the selected conversion method, and coefficient data corresponding to the first conversion method and the second conversion method. Can be generated.

In addition, in the image signal processing unit 110 ′ of FIG. 31, the horizontal and vertical resolution of the image is set by setting a parameter h specifying horizontal resolution and a parameter v specifying vertical resolution, and adjusting these parameters h and v values. Although it is shown that can be adjusted, as described in the part of the TV receiver 100 of FIG. 1, a parameter r for specifying horizontal and vertical resolutions and a parameter z for specifying noise reduction (noise reduction) are set. By adjusting these parameters r and z values, the horizontal and vertical resolutions of the image and the degree of noise reduction can be similarly configured. In this case, in the information memory bank 135 'of the image signal processing unit 110', coefficient data for each combination of class and parameter r and z values corresponding to each conversion method are stored in advance.

This coefficient data can be generated by the coefficient data generating device 180 shown in FIG. 32 similarly to the case of generating coefficient data corresponding to the above-described parameters h and v values. In that case, the parameters r and z are supplied to the SD signal generation circuit 152 as control signals, and when the SD signals are generated from the HD signals corresponding to the values of these parameters r and z, the SD signals are either horizontal or vertical. The band and the noise adding state to the SD signal are varied step by step.

In addition, in the image signal processing unit 110 ′ of FIG. 31, the horizontal and vertical resolution of the image is set by setting a parameter h specifying horizontal resolution and a parameter v specifying vertical resolution, and adjusting these parameters h and v values. Although it is shown that can be adjusted, as described in the part of the television receiver 100 of FIG. 1, the parameters h and v which respectively specify horizontal and vertical resolutions, and the parameters which specify noise reduction (noise reduction), respectively. By setting z and adjusting these parameters h, v, and z values, it is also possible to similarly configure that the horizontal and vertical resolution and noise removal degree of the image can be adjusted. In this case, coefficient data for each combination of class and parameter h, v, and z values corresponding to each conversion method is stored in advance in the information memory bank 135 'of the image signal processing unit 110'.

This coefficient data can also be generated by the coefficient data generating device 180 shown in FIG. 32 similarly to the case of generating coefficient data corresponding to the above-described parameters h and v values. In that case, the parameters h, v, z are supplied to the SD signal generation circuit 152 as a control signal, and when the SD signal is generated from the HD signal in response to these parameters h, v, z values, The noise adding states for the horizontal and vertical bands and the SD signal are varied in stages.

Incidentally, the image signal processing unit 110 ′ of FIG. 31 is processed similarly to the process of the image signal processing unit 110 of FIG. 1 by the image signal processing apparatus 300 shown in FIG. 27, for example. It is also possible to realize this in software. In this case, the coefficient data is stored in advance and used in the ROM 302 or the like.

Referring to the flowchart in FIG. 33, the processing procedure for obtaining the HD signal from the SD signal in the image signal processing device shown in FIG. 27 will be described.

First, processing is started in step ST61, and in step S62, SD pixel data is input in units of frames or fields. When the SD pixel data is input from the input terminal 314, the SD pixel data is temporarily stored in the RAM 303. When the SD pixel data is recorded on the hard disk, the SD pixel data is read by the hard disk drive 305 and temporarily stored in the RAM 303. In step ST63, it is determined whether or not the processing of all frames or all fields of the input SD pixel data is finished. When the process has ended, the process ends in step ST64. On the other hand, when the processing is not finished, the process proceeds to step ST65.

In this step ST65, the image quality specification values (for example, the parameters h and v values, etc.) input by the user by operating the remote control transmitter 200 are read from the RAM 303, for example. In step ST66, the coefficient data Wi of each class corresponding to the image quality specification value is read from the ROM 302 or the like based on the read image quality specification value, and is temporarily stored in the RAM 303.

Next, in step ST67, from the SD pixel data input in step ST62, pixel data of a class tap and a prediction tap are obtained corresponding to each HD pixel data to be generated. Then, in step ST68, it is determined whether or not the process of obtaining the HD pixel data in the entire area of the inputted SD pixel data has ended. When it is finished, the process returns to step ST62 and the process shifts to the input processing of the SD pixel data of the next frame or field. On the other hand, if the processing has not ended, the process proceeds to step ST69.

In step ST69, the class code CL is generated from the SD pixel data of the class tap acquired in step ST67. Then, in step ST70, HD pixel data is generated by the estimation equation using coefficient data corresponding to the class code CL and SD pixel data of the prediction tap, and then the flow returns to step ST67 after that. Repeat the process.

In this manner, by processing according to the flowchart described in FIG. 33, the SD pixel data constituting the input SD signal can be processed to obtain HD pixel data constituting the HD signal. As described above, the HD signal obtained by this processing is output to the output terminal 315, or is supplied to the display 311 to display an image corresponding thereto, or is supplied to the hard disk drive 305 and recorded on the hard disk. Sometimes.

In addition, although illustration of a processing apparatus is abbreviate | omitted, the process in the coefficient data generation apparatus 180 of FIG. 32 can also be implemented by software.

A processing procedure for generating coefficient data will be described with reference to the flowchart in FIG. 34.

First, in step ST81, processing starts, and in step ST82, an image quality pattern (specified by parameters h and v, for example) used for learning is selected. In step ST83, it is determined whether the calculation process of the coefficient data for the entire picture quality pattern has ended. If it is not finished, the flow advances to step ST84.

In this step ST84, known HD pixel data is input in units of frames or fields. Then, in step ST85, it is determined whether or not the processing is finished for all the HD pixel data. When it is not finished, in step ST86, SD pixel data is generated from the HD pixel data input in step ST84 based on the image quality pattern selected in step ST82.

In step ST87, the pixel data of the class tap and the prediction tap are obtained from the SD pixel data generated in step ST86, corresponding to each HD pixel data input in step ST84. In step ST88, it is determined whether or not the learning process has ended in the entire area of the generated SD pixel data. When the learning process is finished, the process returns to step ST84 to input the next HD pixel data, and the same process as described above is repeated. When the learning process is not finished, the process goes to step ST89.

In step ST89, the class code CL is generated from the SD pixel data of the class tap acquired in step ST87. Then, in step ST90, a normal equation (see Equation 21) for obtaining coefficient data is generated. After that, the process returns to step ST87.

When the processing is completed for all the HD pixel data in step ST85 described above, in step ST91, the normal equation generated in step ST90 is solved by a Gaussian-Jordan erasure method or the like to calculate coefficient data of each class. After that, the process returns to step ST82, the next image quality pattern is selected, the same processing as described above is repeated, and coefficient data of each class corresponding to the next image quality pattern is obtained.

When the calculation processing of the coefficient data for the entire image quality pattern is completed in step ST83 described above, in step ST92, the coefficient data of each class for the entire image quality pattern is stored in the memory, and then the processing is performed in step ST93. Quit.

As described above, by performing the processing according to the flowchart illustrated in FIG. 34, coefficient data of each class for the entire picture quality pattern can be obtained by the same method as the coefficient data generation device 180 shown in FIG. 32.

In addition, although the above-mentioned embodiment mentioned the example which used the linear linear equation as an estimation formula at the time of generating an HD signal, it is not limited to this, For example, you may use a higher-order equation as an estimation formula.

In the above-described embodiment, an example of converting an SD signal (525i signal) into an HD signal (525p signal or 1050i signal) is given. However, the present invention is not limited thereto, but the first image is estimated using an estimation equation. It goes without saying that the same applies to the other cases in which the signal is converted into the second image signal.

In addition, in the above-described embodiment, by changing the input parameter value, the functions of resolution enhancement and noise suppression (noise reduction) can be switched continuously. In addition to these resolution enhancement and noise suppression, decoding and signal format conversion are also performed. It is also possible to configure a switchable function as well.

35 shows an image signal processing apparatus 500 capable of switching functions such as resolution enhancement, noise suppression, decoding of MPEG signals, decoding of JPEG (Joint Photographic Experts Group) signals, conversion of composite signals to component signals, and the like. Illustrated.

The image signal processing apparatus 500, the input video signal V _in the input input terminal 501 and, in the process for the input video signal V _in input to the input terminal 501, the output video signal V _out which It consists of the image signal processing part 502 obtained, and the output terminal 503 which outputs the output video signal V _out obtained by this image signal processing part 502. As shown in FIG.

The parameter P is input to the image signal processing unit 502. This parameter P is for selecting the function of the image signal processing unit 502. For example, as shown in Fig. 36, the resolution enhancement function is selected when P = P ₁ , the noise suppression function is selected when P = P ₂ , and the MPEG signal (rate) when P = P _3. the decoding function of b) is selected, P = P ₄ a decoding function of the MPEG signal (acrylate b) is selected when the, P = P ₅ the conversion function of a component signal from the composite signal is selected when the, P = P _{At 6} , the decode function of the JPEG signal is selected. In the image signal processing unit 502, the selected function is processed.

In addition, the image signal processing unit 502 is similar to the image signal processing unit 110 of FIG. 1, and includes a class detection unit 502a that detects a class CL from data of a class tap extracted from the input image signal V _in , and this class. A coefficient data generator 502b for generating coefficient data Wi of the estimated equation corresponding to the class CL and parameter P detected by the detector 502a, and coefficient data generated by the coefficient data generator 502b; It has an input video signal data from the generator for data of the prediction tap extracted from the V _in generating the data which forms an output image signal V _out (502c). The coefficient data generation unit 502b has a memory in which coefficient data Wi of each class corresponding to, for example, a parameter P value, namely P _{1 to} P ₆ , is stored, and is detected by the class detection unit 502a from this memory. The coefficient data Wi corresponding to the received class CL and the input parameter P value is read and output.

In this case, the coefficient data Wi of each class corresponding to the parameter P value, that is, P _{1 to} P ₆ , respectively, is the same as the coefficient data generating device 180 shown in FIG. 32 described above. Can be generated. Here, the part of the SD signal generation circuit 152 in the coefficient data generation device 180 is a student signal generation circuit.

For example, when generating coefficient data corresponding to the parameter P ₁ , a high resolution video signal is input to the input terminal 151 as a teacher signal, and the parameter P ₁ is input to the student signal generation circuit to generate this student signal. In the circuit, a band limiting filter is used from the teacher signal to produce a low resolution video signal as a student signal. Accordingly, the normal equation generating unit 171 generates a normal equation (see Equation 21) for obtaining coefficient data for each class based on the above-described teacher signal and student signal, corresponding to the parameter P ₁ . By solving the equation, coefficient data Wi of each class corresponding to parameter P ₁ for selecting a function of resolution enhancement is obtained.

Also for example, the time to generate the coefficient data corresponding to the parameter P _2, the input video signal as a teacher signal input terminal 151, and by entering the parameter P ₂ in the student signal generating circuit, the student signal generating circuit In Figure 1, noise is added to a video signal as a teacher signal to generate a video signal as a student signal. Accordingly, the normal equation generating unit 171 generates a normal equation (see Equation 21) for obtaining coefficient data for each class based on the above-described teacher signal and student signal, corresponding to the parameter P ₂ . By solving the equation, coefficient data Wi of each class corresponding to parameter P ₂ for selecting the function of noise suppression is obtained.

Also, for example, the time to generate the coefficient data corresponding to the parameter P _3, an input terminal 151 as a teacher signal in the MPEG signal (rate a) input the signal after decoded, and the parameter P ₃ in the student signal generating circuit of the Is input, and the student signal generation circuit generates a signal before decoding the MPEG signal (rate a) from the teacher signal. Accordingly, the normal equation generating unit 171 generates a normal equation (see Equation 21) for obtaining the coefficient data for each class based on the teacher signal and the student signal described above, corresponding to the parameter P ₃ . By solving the normal equation, coefficient data Wi of each class corresponding to parameter P ₃ for selecting the decoding function of the MPEG signal (rate a) is obtained.

Also for example, the time to generate the coefficient data corresponding to parameters P _4, an input terminal 151 to a teacher signal a MPEG signal (rate b) parameters P ₄ to input the signal after decoding, and the student signal generating circuit of the The student signal generation circuit generates a signal before decoding the MPEG signal (rate b) from the teacher signal. Thus, the normal equation generator 171. In the basis of the above teacher signal and the student signal, in response to parameters P _4, the normal equation for obtaining the per-class, the coefficient data (see Equation 21) is created, the By solving the normal equation, coefficient data Wi of each class corresponding to parameter P ₄ for selecting the decoding function of the MPEG signal (rate a) is obtained.

Also for example, the time to generate the coefficient data corresponding to the parameters P _5, the input component signal as a teacher signal input terminal 151, and to input the parameters P ₅ in the student signal generating circuit, the student signal generating circuit In the above, a composite signal as a student signal is generated from the teacher signal. Thus, the normal equation generator 171. In the basis of the above teacher signal and the student signal, in response to the parameters P _5, the normal equation for obtaining the per-class, the coefficient data (see Equation 21) is created, the By solving the normal equation, coefficient data Wi of each class corresponding to parameter P ₅ for selecting the conversion function from the composite signal to the component signal is obtained.

Also for example, the time to generate the coefficient data corresponding to the parameter P _6, input a JPEG signal after the decoding as a teacher signal input terminal 151, and by entering the parameter P ₆ in the student signal generating circuit, the student signal In the generation circuit, a JPEG signal before decoding as a student signal is generated from the teacher signal. Thus, the normal equation generator 171. In the basis of the above teacher signal and the student signal, in response to the parameter P _6, the normal equation for obtaining the per-class, the coefficient data (see Equation 21) is created, the By solving the normal equation, coefficient data Wi of each class corresponding to parameter P ₆ for selecting the decoding function of the JPEG signal is obtained.

In addition, a generation formula for generating coefficient data Wi of the estimation formula as shown in equation (43) is set, and in the memory, coefficient type data w _{0 to} w _n which are coefficient data of the generation formula are stored for each class, and this generation The coefficient data Wi corresponding to the class CL and the input parameter P value detected by the class detection unit 502a may be calculated and obtained by the equation.

In this case, the coefficient seed data w _{0 to} w _n of each class is the same as the coefficient seed data generating device 150 shown in FIG. 18 or the coefficient seed data generating device 150 'shown in FIG. 21 described above. Can be generated using the coefficient seed data generating device. Here, the part of the SD signal generating circuit 152 in the coefficient seed data generating apparatuses 150 and 150 'is used as the student signal generating circuit.

In this case, the parameters P ₁ , P ₂ , P ₃ , P ₄ , P ₆ , and P ₆ are sequentially input to the student signal generation circuit.

When the parameter P ₁ is input, a high resolution video signal is input to the input terminal 151 as a teacher signal, and the student signal generation circuit generates a low resolution video signal as a student signal using a band limiting filter from the teacher signal. do. When inputting the parameter P ₂ , a video signal as a teacher signal is input to the input terminal 151, and noise is added to the video signal as the teacher signal by the student signal generation circuit to generate a video signal as the student signal. When the parameter P ₃ is input, the decoded signal of the MPEG signal (rate a) is input to the input terminal 151 as a teacher signal, and the decode of the MPEG signal (rate a) from the teacher signal is decoded from the teacher signal generation circuit. Generate a signal.

When inputting the parameter P ₄ , the decoded signal of the MPEG signal (rate b) is input to the input terminal 151 as a teacher signal, and the decode of the MPEG signal (rate b) from the teacher signal is decoded from the teacher signal generation circuit. Generate a signal. Further, when the input parameters P _5, the input component signal as a teacher signal input terminal 151, and generates a composite signal as a student signal from the teacher signal at a student signal generating circuit. When inputting the parameter P ₆ , the decoded JPEG signal is input to the input terminal 151 as the teacher signal, and the JPEG signal before decoding as the student signal is generated from the teacher signal by the student signal generation circuit.

Accordingly, in the coefficient seed data generation device corresponding to the coefficient seed data generation device 150 of FIG. 18, in the normal equation generation unit 160, the regular equations for obtaining coefficient seed data w _{0 to} w _n for each class ( (13) is generated, and coefficient type data w _{0 to} w _n of each class are obtained by solving this regular equation.

On the other hand, in the coefficient seed data generation device corresponding to the coefficient seed data generation device 150 'of FIG. 21, in the regular equation generation unit 171, the coefficient data for each class corresponds to each of the parameters P _{1 to} P ₆ . A normal equation (see Equation 21) for obtaining the equation is generated, and by solving this regular equation, coefficient data Wi of each class corresponding to each of the parameters P _{1 to} P ₆ is obtained. In the regular equation generation unit 173, a regular equation for obtaining the coefficient seed data w _{0 to} w _n for each class from coefficient data Wi of each class corresponding to each of the parameters P _{1 to} P ₆ (Equation 26). ), And by solving this regular equation, coefficient class data w _{0 to} w _n of each class are obtained.

The operation of the image signal processing apparatus 500 shown in FIG. 35 will be described.

The input image signal V _in supplied to the input terminal 501 is supplied to the image signal processing unit 502. The input image signal V _in supplied to this image signal processing unit 502 is supplied to the class detection unit 502a. In this class detector 502a, class CL is detected based on the data of the class tap extracted from the input image signal V _in . The class CL detected by the class detector 502a in this way is supplied to the coefficient data generator 502b.

The parameter P input to the image signal processing unit 502 is also supplied to the coefficient data generation unit 502b. In the coefficient data generation unit 502b, coefficient data Wi of the estimation equation corresponding to the class CL and corresponding to the parameter P value is generated. The coefficient data Wi of this estimation formula is supplied to the data generating unit 502c.

In addition, the input image signal Vin is supplied to the data generation unit 502c. In the data generation unit 502c, the data of the prediction tap is extracted from the input image signal V _in , and data constituting the output image signal V _out is generated by the estimation formula using coefficient data Wi. The data generated by this data generation unit 502c is output from the image signal processing unit 502 as an output image signal V _out , and this output image signal V _out is output to the output terminal 503.

As described above, the coefficient data generation unit 502b of the image signal processing unit 502 generates coefficient data Wi corresponding to the input parameter P value, and the data generation unit 502c uses the coefficient data Wi. To generate the data constituting the output image signal V _out . Therefore, in the image signal processing unit 502, the function selected by the parameter P is processed. In other words, by changing the parameter P value, the function of the image signal processing apparatus 500 can be switched.

As described above, according to the image signal processing apparatus 500 shown in FIG. 35, in a single device, functions of resolution enhancement, noise suppression, decoding of MPEG signals, decoding of JPEG signals, and conversion of composite signals to component signals are provided. It can be realized.

In addition, although the resolution of the spatial direction was mentioned as an example of this time as the resolution improvement, the resolution of the temporal direction is also considered. In addition, the function can be switched between one-dimensional Y / C separation, two-dimensional Y / C separation, and three-dimensional Y / C separation.

In the image signal processing apparatus 500 of FIG. 35, the parameter P may be designed to be input by a user, or may be set automatically in accordance with the characteristics of the input video signal V _in . The tap select circuit in the class detector 502a and the predict tap select circuit in the data generator 502c may select each tap according to the parameter P value.

In addition, although the image signal processing circuit 500 of FIG. 35 shows that the parameter P takes discrete values, it is also conceivable that this parameter P takes continuous values. In this case, the coefficient data Wi corresponding to the parameter P can be obtained by linear interpolation using discrete coefficient data or by substituting the parameter P value when using coefficient seed data. Thus, by making the parameter P take a continuous value, for example, as shown in FIG. 36, in decoding the MPEG signals of rates a and b, it is also possible to select an arbitrary rate between these rates a and b. It is also possible to decode the MPEG signal.

In addition, the function which switching is performed in the image signal processing apparatus 500 of FIG. 35 is an example, It is not limited to this. It goes without saying that other functions can be switched by the same configuration.

In the above embodiment, the case where the information signal is an image signal is shown, but the present invention is not limited thereto. For example, even when the information signal is an audio signal, the present invention can be similarly applied.

According to the present invention, the information data constituting the second information signal is generated from the information data constituting the first information signal in response to a parameter value for determining one function among the plurality of functions. Processing of the function can be realized.

According to the present invention, when converting a first information signal to a second information signal, the second information signal is generated corresponding to a plurality of types of parameter values, and the quality of the output obtained by the second information signal is reduced to a plurality of axes. Can be adjusted freely.

As described above, the information signal processing apparatus, the information signal processing method, the image signal processing apparatus, and the image display apparatus using the same according to the present invention, coefficient seed data generating apparatus and generating method, coefficient data generating apparatus and generating method and information used therein The providing medium is suitable for application in improving resolution, suppressing noise, decoding, converting a signal format, or converting an NTSC system video signal to a high-vision video signal.

Claims (27)

delete delete An information signal processing apparatus for converting a first image signal composed of a plurality of information data into a second image signal composed of a plurality of information data, First data selecting means for selecting a plurality of first information data located around the target position in the second image signal based on the first image signal; Class detecting means for detecting a class to which the information data of the target position belongs, based on the plurality of first information data selected by the first data selecting means; Parameter adjusting means for adjusting a plurality of types of parameter values for determining the quality of the output obtained by said second image signal; Information data generating means for generating information data of the point of interest corresponding to a class detected by the class detecting means and a plurality of types of parameter values adjusted by the parameter adjusting means Information signal processing apparatus comprising a. The method of claim 3, The information data generating means, First memory means for storing coefficient type data, which is coefficient data in a generation formula including the plurality of types of parameters for generating coefficient data used in the estimation formula, which are obtained in advance for each class detected by the class detecting means; By the generation formula using the coefficient seed data stored in the first memory means and a plurality of types of parameter values adjusted by the parameter adjusting means, and by the class and the parameter adjusting means detected by the class detecting means. Coefficient data generating means for generating coefficient data of the estimation equation corresponding to the adjusted plurality of types of parameter values; Second data selecting means for selecting a plurality of second information data located around the target position in the second image signal based on the first image signal; Arithmetic means for calculating the information data of the target position based on the estimation equation using the coefficient data generated by the coefficient data generating means and the plurality of second information data selected by the second data selecting means. Information signal processing apparatus comprising a. The method of claim 4, wherein The coefficient data generating means, A coefficient for generating coefficient data of the estimation equation for each class detected by the class detection means by the generation expression, using the coefficient seed data stored in the first memory means and the adjusted plural kinds of parameter values Data generation means, Second memory means for storing coefficient data of the estimation equation in each class generated by the coefficient data generating means; Coefficient data reading means for reading and outputting coefficient data of the estimation equation corresponding to the class detected by the class detecting means from the second memory means Information signal processing apparatus comprising a. The method of claim 4, wherein Addition means for obtaining a total of coefficient data of the estimation equation generated by the coefficient data generating means; Normalization means for dividing and normalizing the information data of the target position obtained by the calculation means by the total. Information signal processing apparatus further comprising a. The method of claim 3, The information data generating means, A memory for storing coefficient data of an estimated expression generated in advance for each combination of a class detected by the class detecting means and a plurality of types of parameter values adjusted by the parameter adjusting means, the class detected by the class detecting means and the parameter Coefficient data generating means for generating coefficient data of the estimation equation corresponding to the plurality of types of parameter values adjusted by the adjusting means; Second data selecting means for selecting a plurality of second information data located around the target position in the second image signal based on the first image signal; Arithmetic means for calculating the information data of the target position based on the estimation equation using the coefficient data generated by the coefficient data generating means and the plurality of second information data selected by the second data selecting means. Information signal processing apparatus comprising a. The method of claim 7, wherein The coefficient data generating means, A first memory unit for storing coefficient data of the estimated equation generated in advance for each combination of a class detected by the class detecting means and a plurality of types of parameter values adjusted by the parameter adjusting means; First data reading means for reading coefficient data of each class corresponding to a plurality of types of parameter values adjusted by the parameter adjusting means from the first memory section; A second memory unit for storing coefficient data of each class read by said first data reading means; Second data reading means for reading coefficient data corresponding to the class detected by the class detecting means from the second memory section; Information signal processing apparatus comprising a. The method of claim 3, The parameter adjusting means, Display means for displaying adjustment positions of the plurality of types of parameters; User operation means for adjusting the plurality of types of parameter values with reference to the display of the display means Information signal processing apparatus comprising a. An image signal processing apparatus for converting a first image signal composed of a plurality of pixel data into a second image signal composed of a plurality of pixel data, Data selecting means for selecting a plurality of pixel data located around the target position in the second image signal based on the first image signal; Class detecting means for detecting a class to which the pixel data of the target position belongs, based on the plurality of pixel data selected by the data selecting means; Parameter adjusting means for adjusting a plurality of types of parameter values for determining the quality of the output obtained by said second image signal; Pixel data generating means for generating pixel data of the target position corresponding to a class detected by the class detecting means and a plurality of types of parameter values adjusted by the parameter adjusting means. Image signal processing apparatus comprising a. Image signal input means for inputting a first image signal composed of a plurality of pixel data; Image signal processing means for converting the first image signal input from the image signal input means into a second image signal composed of a plurality of pixel data and outputting the second image signal; Image display means for displaying an image by said second image signal output from said image signal processing means on an image display element; Parameter adjusting means for adjusting a plurality of types of parameter values for determining image quality of the image displayed on the image display element Including; The image signal processing means, Data selecting means for selecting a plurality of pixel data located around the target position in the second image signal based on the first image signal; Class detecting means for detecting a class to which the pixel data of the target position belongs, based on the plurality of pixel data selected by the data selecting means; Pixel data generating means for generating pixel data of the target position corresponding to a class detected by the class detecting means and a plurality of types of parameter values adjusted by the parameter adjusting means. Image display device comprising a. The method of claim 11, The parameter adjusting means, Display control means for displaying the adjustment positions of the plurality of types of parameters on the image display element; User operation means for adjusting the plurality of types of parameter values with reference to the adjustment positions of the plurality of types of parameters displayed on the image display element Image display device comprising a. An information signal processing method for converting a first image signal composed of a plurality of information data into a second image signal composed of a plurality of information data, A first step of selecting a plurality of first information data located around the target position in the second image signal based on the first image signal; A second step of detecting a class to which the information data of the target location belongs, based on the plurality of first information data selected in the first step; A third step of adjusting a plurality of types of parameter values for determining the quality of the output obtained by said second image signal, A fourth step of generating information data of the point of interest corresponding to the class detected in the second step and the plurality of types of parameter values adjusted in the third step; Information signal processing method comprising a. The method of claim 13, The fourth step, Generating coefficient data of an estimation equation corresponding to the class detected in the second step and the plurality of types of parameter values adjusted in the third step; Selecting a plurality of second information data located around the target position in the second image signal based on the first image signal; Calculating the information data of the point of interest based on the estimation equation using the generated coefficient data and the plurality of second information data. Information signal processing method comprising a. The method of claim 14, In the step of generating the coefficient data, Coefficient seed data, which is coefficient data in a generation formula including the plurality of types of parameters for generating coefficient data used in the estimation formula, obtained in advance for each class detected in the second stage, and a plurality of kinds of adjusted in the third stage. Calculating coefficient data of the estimation equation corresponding to the class detected in the second step and the plurality of types of parameter values adjusted in the third step by using the parameter value. Information signal processing method. The method of claim 15, A fifth step of obtaining a sum of coefficient data of the estimation equation generated in the step of generating the coefficient data; A sixth step of dividing and normalizing the information data of the target position obtained in the fourth step by the total obtained in the fifth step; Information signal processing method further comprising. The method of claim 14, In the step of generating the coefficient data, The class detected in the second step and the third step from a storage unit in which coefficient data of the estimation equation for each combination of the class detected in the second step and the plurality of types of parameter values adjusted in the third step is stored And based on the plurality of types of parameter values adjusted in step S, the coefficient data of the estimation equation corresponding to the class detected in the second step and the plurality of types of parameter values adjusted in the third step are read and obtained. Information signal processing method. In order to convert the first image signal composed of a plurality of information data into a second image signal composed of a plurality of information data, A first step of selecting a plurality of information data located around a target position in the second image signal based on the first image signal; A second step of detecting a class to which the information data of the target location belongs, based on the plurality of information data selected in the first step; A third step of adjusting a plurality of types of parameter values for determining the quality of the output obtained by said second image signal, A fourth step of generating information data of the point of interest corresponding to the class detected in the second step and the plurality of types of parameter values adjusted in the third step; An information providing medium for providing a computer program for executing the program. Apparatus for generating coefficient seed data which is coefficient data in a generation equation for generating coefficient data used in an estimation equation used when converting a first image signal composed of a plurality of information data into a second image signal composed of a plurality of information data as, Signal processing means for processing a teacher signal corresponding to the second image signal to obtain an input signal corresponding to the first image signal; Parameter adjusting means for adjusting a plurality of types of parameter values for determining the quality of an output obtained by the input signal, corresponding to the plurality of types of parameters included in the generation formula; First data selecting means for selecting a plurality of first information data located around the target position in the teacher signal based on the input signal; Class detecting means for detecting a class to which the information data of the target position belongs, based on the plurality of first information data selected by the first data selecting means; Second data selecting means for selecting a plurality of second information data located around the target position in the teacher signal based on the input signal; Normal for obtaining the coefficient seed data for each class, using the class detected by the class detecting means, the plurality of second information data selected by the second data selecting means, and the information data of the target position in the teacher signal. Regular equation generating means for generating equations, Coefficient seed data computing means for solving the regular equations and obtaining the coefficient seed data for each class Coefficient count data generating apparatus comprising a. A method of generating coefficient seed data which is coefficient data in a generation formula for generating coefficient data used in an estimation equation used when converting a first image signal composed of a plurality of information data into a second image signal composed of a plurality of information data As A first step of processing a teacher signal corresponding to the second image signal to obtain an input signal corresponding to the first image signal; A second step of adjusting a plurality of types of parameter values for determining the quality of an output obtained by the input signal, corresponding to the plurality of types of parameters included in the generation formula; A third step of selecting, from the input signal, a plurality of first information data located around the target position in the teacher signal; A fourth step of detecting a class to which the information data of the target location belongs, based on the plurality of first information data selected in the third step; A fifth step of selecting a plurality of second information data located around the target position in the teacher signal based on the input signal; Using the class detected in the fourth step, the plurality of second information data selected in the fifth step, and the information data of the target position in the teacher signal, a regular equation for obtaining the coefficient seed data is obtained for each class. The sixth step of generating, A seventh step of solving the regular equation and obtaining the coefficient seed data for each class Counting species data generation method comprising a. To generate coefficient seed data which is coefficient data in a generation formula for generating coefficient data used in an estimation equation used when converting a first image signal composed of a plurality of information data into a second image signal composed of a plurality of information data. , A first step of processing a teacher signal corresponding to the second image signal to obtain an input signal corresponding to the first image signal; A second step of adjusting a plurality of types of parameter values for determining the quality of an output obtained by the input signal, corresponding to the plurality of types of parameters included in the generation formula; A third step of selecting a plurality of first information data located around the target position in the teacher signal based on the input signal; A fourth step of detecting a class to which the information data of the target location belongs, based on the plurality of first information data selected in the third step; A fifth step of selecting a plurality of second information data located around the target position in the teacher signal based on the input signal; Using the class detected in the fourth step, the plurality of second information data selected in the fifth step, and the information data of the target position in the teacher signal, a regular equation for obtaining the coefficient seed data is obtained for each class. The sixth step of generating, A seventh step of solving the regular equation and obtaining the coefficient seed data for each class An information providing medium for providing a computer program for executing the program. Apparatus for generating coefficient seed data which is coefficient data in a generation formula for generating coefficient data used in an estimation equation used when converting a first image signal composed of a plurality of information data into a second image signal composed of a plurality of information data as, Signal processing means for processing a teacher signal corresponding to the second image signal to obtain an input signal corresponding to the first image signal; Parameter adjusting means for adjusting a plurality of types of parameter values for determining the quality of an output obtained by the input signal, corresponding to the plurality of types of parameters included in the generation formula; First data selecting means for selecting a plurality of first information data located around the target position in the teacher signal based on the input signal; Class detecting means for detecting a class to which the information data of the target position belongs, based on the plurality of first information data selected by the first data selecting means; Second data selecting means for selecting a plurality of second information data located around the target position in the teacher signal based on the input signal; A combination of the class and the plurality of types of parameter values by using a class detected by the class detecting means, the plurality of second information data selected by the second data selecting means, and information data of a target position in the teacher signal. First normal equation generating means for generating a first normal equation for obtaining coefficient data of the estimated equation each time; Coefficient data calculating means for solving the first normal equation and obtaining coefficient data of the estimated equation for each combination; Second normal equation generating means for generating a second normal equation for obtaining the coefficient seed data for each class using coefficient data for each combination obtained by the coefficient data calculating means; Coefficient seed data calculating means for solving the second normal equation and obtaining the coefficient seed data for each class Coefficient count data generating apparatus comprising a. A method of generating coefficient seed data which is coefficient data in a generation formula for generating coefficient data used in an estimation equation used when converting a first image signal composed of a plurality of information data into a second image signal composed of a plurality of information data As A first step of processing a teacher signal corresponding to the second image signal to obtain an input signal corresponding to the first image signal; A second step of adjusting a plurality of types of parameter values for determining the quality of an output obtained by the input signal, corresponding to the plurality of types of parameters included in the generation formula; A third step of selecting a plurality of first information data located around the target position in the teacher signal based on the input signal; A fourth step of detecting a class to which the information data of the target location belongs, based on the plurality of first information data selected in the third step; A fifth step of selecting a plurality of second information data located around the target position in the teacher signal based on the input signal; In the estimation equation for each combination of a class and a plurality of types of parameter values, using the class detected in the fourth step, the plurality of second information data selected in the fifth step and the information data of the target position in the teacher signal. A sixth step of generating a first normal equation for obtaining coefficient data to be used; A seventh step of solving the first normal equation and obtaining coefficient data of the estimated equation for each combination; An eighth step of generating, by class, a second normal equation for obtaining the coefficient seed data using the coefficient data for each combination obtained in the seventh step; A ninth step of solving the second normal equation to obtain the coefficient seed data for each class Counting species data generation method comprising a. To generate coefficient seed data which is coefficient data in a generation formula for generating coefficient data used in an estimation equation used when converting a first image signal composed of a plurality of information data into a second image signal composed of a plurality of information data. , A first step of processing a teacher signal corresponding to the second image signal to obtain an input signal corresponding to the first image signal; A second step of adjusting a plurality of types of parameter values for determining the quality of an output obtained by the input signal, corresponding to the plurality of types of parameters included in the generation formula; A third step of selecting a plurality of first information data located around the target position in the teacher signal based on the input signal; A fourth step of detecting a class to which the information data of the target location belongs, based on the plurality of first information data selected in the third step; A fifth step of selecting a plurality of second information data located around the target position in the teacher signal based on the input signal; The estimation equation for each combination of class and plural kinds of parameter values using the class detected in the fourth step, the plurality of second information data selected in the fifth step, and information data of a target position in the teacher signal. Generating a first regular equation for obtaining coefficient data of; A seventh step of solving the first normal equation and obtaining coefficient data of the estimated equation for each combination; An eighth step of generating a second normal equation for obtaining the coefficient seed data for each class from the coefficient data for each combination obtained in the seventh step; A ninth step of solving the second normal equation to obtain the coefficient seed data for each class An information providing medium for providing a computer program for executing the program. An apparatus for generating coefficient data of an estimation equation used when converting a first image signal composed of a plurality of information data into a second image signal composed of a plurality of information data, Signal processing means for processing a teacher signal corresponding to the second image signal to obtain an input signal corresponding to the first image signal; Parameter adjusting means for adjusting a plurality of types of parameter values for determining the quality of the output obtained by the input signal; First data selecting means for selecting a plurality of first information data located around the target position in the teacher signal based on the input signal; Class detecting means for detecting a class to which the information data of the target position belongs, based on the plurality of first information data selected by the first data selecting means; Second data selecting means for selecting a plurality of second information data located around the target position in the teacher signal based on the input signal; A combination of the class and the plurality of types of parameter values by using a class detected by the class detecting means, the plurality of second information data selected by the second data selecting means, and information data of a target position in the teacher signal. Regular equation generating means for generating a regular equation for obtaining coefficient data of the estimated equation for each; Coefficient data calculating means for solving the regular equation to obtain the coefficient data for each combination Coefficient data generation apparatus comprising a. A method of generating coefficient data of an estimation equation used when converting a first image signal composed of a plurality of information data into a second image signal composed of a plurality of information data, A first step of processing a teacher signal corresponding to the second image signal to obtain an input signal corresponding to the first image signal; A second step of adjusting a plurality of types of parameter values for determining the quality of the output obtained by the input signal; A third step of selecting a plurality of first information data located around the target position in the teacher signal based on the input signal; A fourth step of detecting a class to which the information data of the target location belongs, based on the plurality of first information data selected in the third step; A fifth step of selecting a plurality of second information data located around the target position in the teacher signal based on the input signal; For each combination of the class and the plurality of types of parameter values using the class detected in the fourth step, the plurality of second information data selected in the fifth step, and information data of the target position in the teacher signal. Generating a normal equation for obtaining coefficient data of the estimation equation; A seventh step of obtaining the coefficient data for each combination by solving the regular equation generated in the sixth step Coefficient data generation method comprising a. In order to generate coefficient data of an estimation equation used when converting a first image signal composed of a plurality of information data into a second image signal composed of a plurality of information data, A first step of processing a teacher signal corresponding to the second image signal to obtain an input signal corresponding to the first image signal; A second step of adjusting a plurality of types of parameter values for determining the quality of the output obtained by the input signal; A third step of selecting a plurality of first information data located around the target position in the teacher signal based on the input signal; A fourth step of detecting a class to which the information data of the target location belongs, based on the plurality of first information data selected in the third step; A fifth step of selecting a plurality of second information data located around the target position in the teacher signal based on the input signal; For each combination of the class and the plurality of types of parameter values using the class detected in the fourth step, the plurality of second information data selected in the fifth step, and information data of the target position in the teacher signal. Generating a normal equation for obtaining coefficient data of the estimation equation; A seventh step of obtaining the coefficient data for each combination by solving the regular equation generated in the sixth step An information providing medium for providing a computer program for executing the program.

Images